TWI846456B - Data analysis method and data analysis device - Google Patents

Abstract

本發明提供一種資料分析方法及資料分析裝置，資料分析方法包含下列步驟：取得一資料集；對該些資料設定一搜尋範圍；對該些資料於該搜尋範圍中搜尋一第一分割線與一第二分割線，該第一分割線與該第二分割線相交於一分割原點；依據該分割原點的該第一分割線與該第二分割線將該些資料區分為四個象限；依據各該象限中的資料數量與相鄰的象限中的資料數量的比值判斷該些資料的一資料關係，並輸出該資料關係。資料分析裝置用以執行上述資料分析方法。The present invention provides a data analysis method and a data analysis device, the data analysis method comprising the following steps: obtaining a data set; setting a search range for the data; searching for a first dividing line and a second dividing line in the search range for the data, the first dividing line and the second dividing line intersecting at a dividing origin; dividing the data into four quadrants according to the first dividing line and the second dividing line at the dividing origin; judging a data relationship of the data according to the ratio of the amount of data in each quadrant to the amount of data in the adjacent quadrant, and outputting the data relationship. The data analysis device is used to execute the above data analysis method.

Description

資料分析方法及資料分析裝置Data analysis method and data analysis device

本發明係與資料分析有關；特別是指一種資料分析方法及資料分析裝置。The present invention relates to data analysis; in particular, it relates to a data analysis method and a data analysis device.

隨著科技的發展，資料的收集已經是很容易的事，無論是工業、農業、醫療、商業等領域，都能夠收集大量的資料。進一步，可藉由收集到的資料進行大數據分析以及人工智慧的應用。With the development of technology, data collection has become very easy. Whether in industry, agriculture, medicine, or business, a large amount of data can be collected. Furthermore, the collected data can be used for big data analysis and artificial intelligence applications.

傳統的資料分析方式通常使用多種統計方法的組合對資料集進行分析，以建立應用模型，分析後的結果多為分類、預測、決策等。統計方法例如是相關係數（correlation coefficient）、廻歸分析、T檢定、單因子單變量變異數分析（ANOVA）、K近鄰演算法、支援向量機等。由於使用多種統計方法，因此，資料分析的運算較複雜且運算量較大，此外，資料分析通常由大型主機或雲端伺服以特定的軟體執行，無法在小型的運算裝置執行。Traditional data analysis methods usually use a combination of multiple statistical methods to analyze data sets in order to establish application models. The results of the analysis are mostly classification, prediction, decision-making, etc. Statistical methods include correlation coefficient, regression analysis, T test, single-factor single-variable analysis of variance (ANOVA), K nearest neighbor algorithm, support vector machine, etc. Due to the use of multiple statistical methods, the data analysis operations are more complex and the amount of calculation is large. In addition, data analysis is usually performed by large mainframes or cloud servers with specific software and cannot be performed on small computing devices.

有鑑於此，本發明之目的在於提供一種資料分析方法及資料分析裝置，可判斷複數個資料的資料關係。In view of this, an object of the present invention is to provide a data analysis method and a data analysis device that can determine the data relationship of a plurality of data.

緣以達成上述目的，本發明提供的一種資料分析方法，包含下列步驟：In order to achieve the above-mentioned purpose, the present invention provides a data analysis method comprising the following steps:

A. 取得一資料集，該資料集包括複數個資料，各該資料包括一第一值與一第二值；A. obtaining a data set, the data set comprising a plurality of data, each of the data comprising a first value and a second value;

B. 對該些資料設定一搜尋範圍，其中，該搜尋範圍具有一下限第一值與一上限第一值，一下限第二值與一上限第二值；B. setting a search range for the data, wherein the search range has a lower limit first value and an upper limit first value, a lower limit second value and an upper limit second value;

C. 對該些資料於該搜尋範圍中搜尋至少一第一分割線與至少一第二分割線，該至少一第一分割線與該至少一第二分割線相交於至少一分割原點，該至少一分割原點具有一原點第一值與一原點第二值，該原點第一值介於該下限第一值與該上限第一值之間，該原點第二值介於該下限第二值與該上限第二值之間，該原點第一值與該原點第二值分別對應該至少一第一分割線與該至少一第二分割線；C. searching for at least one first cutting line and at least one second cutting line in the search range for the data, the at least one first cutting line and the at least one second cutting line intersecting at at least one cutting origin, the at least one cutting origin having a first origin value and a second origin value, the first origin value being between the first lower limit value and the first upper limit value, the second origin value being between the second lower limit value and the second upper limit value, the first origin value and the second origin value corresponding to the at least one first cutting line and the at least one second cutting line respectively;

D. 依據該至少一分割原點的該至少一第一分割線與該至少一第二分割線將該些資料區分為四個象限，該些象限中分別具有該些資料的一部分；D. dividing the data into four quadrants according to the at least one first dividing line and the at least one second dividing line of the at least one dividing origin, wherein the quadrants respectively contain a portion of the data;

E. 依據基於該至少一分割原點的各該象限中的資料數量與相鄰的象限中的資料數量的比值判斷該些資料的一資料關係，並輸出基於該至少一分割原點的該資料關係。E. determining a data relationship of the data based on a ratio of the data quantity in each quadrant based on the at least one segmentation origin to the data quantity in an adjacent quadrant, and outputting the data relationship based on the at least one segmentation origin.

本發明提供的一種資料分析裝置，包含一通訊單元與一處理單元，其中，該通訊單元供連接一通訊介面且自該通訊介面接收一資料集，該資料集包括複數個資料，各該資料包括一第一值與一第二值；該處理單元電性連接該通訊單元，該處理單元取得該資料集，並且對該些資料於一搜尋範圍中搜尋至少一第一分割線與至少一第二分割線，其中，該搜尋範圍具有一下限第一值與一上限第一值，一下限第二值與一上限第二值；該至少一第一分割線與該至少一第二分割線相交於至少一分割原點，該至少一分割原點具有一原點第一值與一原點第二值，該原點第一值介於該下限第一值與該上限第一值之間，該原點第二值介於該下限第二值與該上限第二值之間；The present invention provides a data analysis device, comprising a communication unit and a processing unit, wherein the communication unit is connected to a communication interface and receives a data set from the communication interface, the data set includes a plurality of data, each of the data includes a first value and a second value; the processing unit is electrically connected to the communication unit, the processing unit obtains the data set, and searches for at least one first dividing line and at least one second dividing line in a search range for the data, wherein the search range has a lower limit first value and an upper limit first value, a lower limit second value and an upper limit second value; the at least one first dividing line and the at least one second dividing line intersect at at least one dividing origin, the at least one dividing origin has an origin first value and an origin second value, the origin first value is between the lower limit first value and the upper limit first value, and the origin second value is between the lower limit second value and the upper limit second value;

其中，該處理單元依據該至少一第一分割線與該至少一第二分割線將該些資料區分為四個象限，該些象限中分別具有該些資料的一部分；該處理單元依據基於該至少一分割原點的各該象限中的資料數量與相鄰的象限中的資料數量的比值判斷該些資料的一資料關係，並輸出基於該至少一分割原點的該資料關係至該通訊單元；The processing unit divides the data into four quadrants according to the at least one first dividing line and the at least one second dividing line, and each of the quadrants contains a portion of the data; the processing unit determines a data relationship of the data according to a ratio between the amount of data in each quadrant based on the at least one dividing origin and the amount of data in an adjacent quadrant, and outputs the data relationship based on the at least one dividing origin to the communication unit;

其中，該通訊單元將基於該至少一分割原點的該資料關係輸出至該通訊介面。The communication unit outputs the data relationship based on the at least one segmentation origin to the communication interface.

本發明之效果在於，藉由將該些資料集的資料以象限分割的方式判斷資料的資料關係，並輸出到外部裝置，可供後續的利用，有別於習用的資料分析方式。另外，本發明的資料分析方法由於不使用複雜的統計方法，除了可用於大型主機、雲端伺服器之外，亦可應用於小型的運算裝置。The effect of the present invention is that the data relationship of the data is determined by dividing the data sets into quadrants and outputting them to an external device for subsequent use, which is different from the conventional data analysis method. In addition, the data analysis method of the present invention does not use complex statistical methods, so it can be used in large mainframes and cloud servers as well as small computing devices.

為能更清楚地說明本發明，茲舉較佳實施例並配合圖式詳細說明如後。請參圖1所示，為本發明第一較佳實施例之資料分析方法所應用的資料分析裝置1，該資料分析裝置1可為一個獨立的裝置、整合於一運算系統（例如電腦或伺服器）中、或為可插拔的硬體裝置（例如Dongle）。In order to more clearly explain the present invention, a preferred embodiment is given and described in detail with reference to the drawings. Please refer to FIG1 , which shows a data analysis device 1 used in the data analysis method of the first preferred embodiment of the present invention. The data analysis device 1 can be an independent device, integrated into a computing system (such as a computer or server), or a pluggable hardware device (such as a dongle).

該資料分析裝置1包含一通訊單元10與一處理單元20。該通訊單元10供連接一通訊介面10a，該通訊單元10透過該通訊介面10a與一外部裝置100通訊，以接收或傳輸資料及/或命令，通訊介面10a例如是網路介面、USB介面等。The data analysis device 1 includes a communication unit 10 and a processing unit 20. The communication unit 10 is connected to a communication interface 10a. The communication unit 10 communicates with an external device 100 through the communication interface 10a to receive or transmit data and/or commands. The communication interface 10a is, for example, a network interface, a USB interface, etc.

該處理單元20電性連接該通訊單元10，以取得該通訊單元10傳來的資料，以及將資料進行分析及/或處理後的結果傳送到該通訊單元10。該處理單元20可例如是MCU、GPU、或CPU等處理器。The processing unit 20 is electrically connected to the communication unit 10 to obtain data transmitted from the communication unit 10 and transmit the result of analyzing and/or processing the data to the communication unit 10. The processing unit 20 may be a processor such as an MCU, a GPU, or a CPU.

本實施例中，該資料分析裝置1可選擇地包含一儲存單元30，該儲存單元30分別電性連接該通訊單元10與該處理單元20。該儲存單元可例如為硬碟、或快閃記憶體等可作為通訊單元10與處理單元20之資料的儲存媒介。該通訊單元10傳來的資料可先儲存在該儲存單元30，再由該處理單元20自儲存單元30讀取，或者由該通訊單元10直接將資料傳輸給該處理單元20。該處理單元20所分析、處理的資料的結果可儲存在該儲存單元30及/或直接輸出給通訊單元10。In this embodiment, the data analysis device 1 may optionally include a storage unit 30, and the storage unit 30 is electrically connected to the communication unit 10 and the processing unit 20, respectively. The storage unit may be, for example, a hard disk or a flash memory, etc., which can be used as a storage medium for the data of the communication unit 10 and the processing unit 20. The data transmitted from the communication unit 10 can be first stored in the storage unit 30, and then read from the storage unit 30 by the processing unit 20, or the communication unit 10 directly transmits the data to the processing unit 20. The results of the data analyzed and processed by the processing unit 20 can be stored in the storage unit 30 and/or directly output to the communication unit 10.

藉由上述之資料分析裝置1，即可進行本實施例的資料分析方法，該方法包含圖2所示之下列步驟：The data analysis method of this embodiment can be performed by using the data analysis device 1, and the method includes the following steps as shown in FIG. 2:

步驟S11：資料分析裝置1取得一資料集，該資料集包括複數個資料，各該資料包括一第一值與一第二值。Step S11: The data analysis device 1 obtains a data set, which includes a plurality of data, each of which includes a first value and a second value.

本實施例中，係由該通訊單元10透過該通訊介面自外部裝置接收該資料集。請配合圖3，該資料集包括複數個資料，各該資料包括一第一值與一第二值。更詳而言，各該資料為二維的資料，各該資料的第一值可例如為X值，第二值可例如為Y值，各該資料的第一值與第二值可在由一第一軸（X軸）與一第二軸Y（Y軸）構成的二維座標系上成為一個點。該些資料可在二維座標系中形成複數個點的分布。實務上，資料集的資料可是應用在任何領域的二維資料，包含工業、農業、醫療、商業...等領域，例如資料集的資料可為從感測器收集的資料，其第一值為時間，第二值為感測數值，但不以此為限。In this embodiment, the communication unit 10 receives the data set from the external device through the communication interface. Referring to FIG. 3 , the data set includes a plurality of data, each of which includes a first value and a second value. More specifically, each of the data is two-dimensional data, and the first value of each of the data may be, for example, an X value, and the second value may be, for example, a Y value. The first value and the second value of each of the data may form a point on a two-dimensional coordinate system formed by a first axis (X axis) and a second axis Y (Y axis). The data may form a distribution of a plurality of points in the two-dimensional coordinate system. In practice, the data in the dataset can be two-dimensional data applied in any field, including industry, agriculture, medicine, commerce, etc. For example, the data in the dataset can be data collected from a sensor, wherein the first value is time and the second value is a sensing value, but not limited thereto.

該通訊單元10取得該資料集後，可將該資料集直接傳送給該處理單元20，或先儲存於該儲存單元30。再由該處理單元20取得該資料集。本實施例中，該處理單元20接收來該通訊單元10所傳送的該資料集，或自該儲存單元30讀取該通訊單元10儲存在該儲存單元30的資料集。After the communication unit 10 obtains the data set, it can directly transmit the data set to the processing unit 20, or store it in the storage unit 30 first. Then the processing unit 20 obtains the data set. In this embodiment, the processing unit 20 receives the data set transmitted by the communication unit 10, or reads the data set stored in the storage unit 30 by the communication unit 10 from the storage unit 30.

步驟S12：資料分析裝置1對該些資料設定一搜尋範圍S，其中，該搜尋範圍S具有一下限第一值與一上限第一值，一下限第二值與一上限第二值。Step S12: The data analysis device 1 sets a search range S for the data, wherein the search range S has a lower limit first value and an upper limit first value, a lower limit second value and an upper limit second value.

本實施例中，該處理單元20在分析該資料集之前，係先設定搜尋範圍S，意即欲分析的資料範圍，供後續對該資料集中第一值介於該下限第一值與該上限第一值且第二值介於該下限第二值與該上限第二值的複數個資料進行分析。該搜尋範圍S可由外部裝置100以命令經由通訊介面10a傳送到該通訊單元10，再由該通訊單元10傳送給該處理單元20。In this embodiment, the processing unit 20 sets a search range S, i.e., a data range to be analyzed, before analyzing the data set, for subsequent analysis of a plurality of data in the data set whose first value is between the lower limit first value and the upper limit first value and whose second value is between the lower limit second value and the upper limit second value. The search range S can be transmitted from the external device 100 to the communication unit 10 via the communication interface 10a as a command, and then transmitted from the communication unit 10 to the processing unit 20.

此外，該搜尋範圍S更可具有一第一搜尋間隔值與一第二搜尋間隔值，第一搜尋間隔值與第二搜尋間隔值用以在步驟S13中使用。舉例而言，於該第一軸X上，該下限第一值為0，該上限第一值為1，第一搜尋間隔值為0.1；於該第二軸Y上，該下限第二值為0，該上限第二值為1，第二搜尋間隔值為0.1。In addition, the search range S may further have a first search interval value and a second search interval value, and the first search interval value and the second search interval value are used in step S13. For example, on the first axis X, the lower limit first value is 0, the upper limit first value is 1, and the first search interval value is 0.1; on the second axis Y, the lower limit second value is 0, the upper limit second value is 1, and the second search interval value is 0.1.

步驟S13：資料分析裝置1對該些資料於該搜尋範圍S中搜尋至少一第一分割線L1與至少一第二分割線L2（圖3參照），該至少一第一分割線L1與該至少一第二分割線L2相交於至少一分割原點Os。Step S13: The data analysis device 1 searches for at least one first cutting line L1 and at least one second cutting line L2 in the search range S (see FIG. 3 ), and the at least one first cutting line L1 and the at least one second cutting line L2 intersect at at least one cutting origin Os.

本實施例中，係由該處理單元20於該搜尋範圍S中對資料進行第一分割線L1與第二分割線L2的搜尋。該第一分割線L1與該第二分割線L2相交於一分割原點Os，該分割原點Os具有一原點第一值（以0.5為例）與一原點第二值（以0.3為例），該原點第一值介於該下限第一值與該上限第一值之間，該原點第二值介於該下限第二值與該上限第二值之間。該原點第一值與該原點第二值分別對應該第一分割線L1與該第二分割線L2。In this embodiment, the processing unit 20 searches for the first cutting line L1 and the second cutting line L2 in the search range S. The first cutting line L1 and the second cutting line L2 intersect at a cutting origin Os, and the cutting origin Os has a first origin value (for example, 0.5) and a second origin value (for example, 0.3). The first origin value is between the first lower limit value and the first upper limit value, and the second origin value is between the second lower limit value and the second upper limit value. The first origin value and the second origin value correspond to the first cutting line L1 and the second cutting line L2, respectively.

本實施例中，處理單元20搜尋該第一分割線L1與該第二分割線L2的步驟包含：In this embodiment, the step of the processing unit 20 searching for the first cutting line L1 and the second cutting line L2 includes:

步驟S131：如圖4所示，設定複數個搜尋原點O，各該搜尋原點O具有一預定第一值與一預定第二值，各該預定第一值介於該下限第一值與該上限第一值之間，各該預定第二值介於該下限第二值與該上限第二值之間。Step S131: As shown in FIG. 4 , a plurality of search origins O are set, each of which has a predetermined first value and a predetermined second value, each of which is between the lower limit first value and the upper limit first value, and each of which is between the lower limit second value and the upper limit second value.

如圖4所示，在該下限第一值至該上限第一值之間，每隔一個第一搜尋間隔值設定一個預定第一值，在該下限第二值至該上限第二值之間，每隔一個第二搜尋間隔值設定一個預定第二值。較佳者，最小的預定第一值不等於下限第一值，最大的預定第一值不等於上限第一值；最小的預定第二值不等於下限第二值，最大的預定第二值不等於上限第二值。各該預定第一值與各該預定第二值形成各該搜尋原點O的座標。圖4中以81個搜尋原點O為例。As shown in FIG4 , between the lower limit first value and the upper limit first value, a predetermined first value is set for every first search interval value, and between the lower limit second value and the upper limit second value, a predetermined second value is set for every second search interval value. Preferably, the smallest predetermined first value is not equal to the lower limit first value, and the largest predetermined first value is not equal to the upper limit first value; the smallest predetermined second value is not equal to the lower limit second value, and the largest predetermined second value is not equal to the upper limit second value. Each predetermined first value and each predetermined second value form the coordinates of each search origin O. FIG4 takes 81 search origins O as an example.

步驟S132：以各該搜尋原點O的預定第一值與預定第二值分別形成的一第一預定分割線與一第二預定分割線將該些資料區分為四個預定象限，該四個預定象限分別為一第一預定象限、一第二預定象限、一第三預定象限與一第四預定象限。Step S132: Divide the data into four predetermined quadrants using a first predetermined dividing line and a second predetermined dividing line formed by the predetermined first value and the predetermined second value of each search origin O. The four predetermined quadrants are a first predetermined quadrant, a second predetermined quadrant, a third predetermined quadrant and a fourth predetermined quadrant.

如圖5為基於搜尋原點O(預定第一值=0.1，預定第二值=0.1)形成的第一預定分割線Lp1與第二預定分割線Lp2所區分成的第一至第四預定象限Qp1~Qp4。As shown in FIG. 5 , the first to fourth predetermined quadrants Qp1 - Qp4 are divided by the first predetermined dividing line Lp1 and the second predetermined dividing line Lp2 formed based on the search origin O (the predetermined first value=0.1, the predetermined second value=0.1).

步驟S132：分別計算基於各該搜尋原點的Z1y、Z2y、Z3y、Z4y、Z5y、Z6y、Z1x、Z2x、Z3x、Z4x、Z5x與Z6x數值。較佳者，以四個預定象限皆有資料的搜尋原點計算Z1y、Z2y、Z3y、Z4y、Z5y、Z6y、Z1x、Z2x、Z3x、Z4x、Z5x與Z6x數值。Step S132: Calculate the values of Z1y, Z2y, Z3y, Z4y, Z5y, Z6y, Z1x, Z2x, Z3x, Z4x, Z5x and Z6x based on the search origins. Preferably, the values of Z1y, Z2y, Z3y, Z4y, Z5y, Z6y, Z1x, Z2x, Z3x, Z4x, Z5x and Z6x are calculated based on the search origins having data in all four predetermined quadrants.

其中：in:

Z1y為C1y與C2y之中的最小值，亦即，Z1y=min(C1y,C2y)；Z1y is the minimum value between C1y and C2y, that is, Z1y=min(C1y,C2y);

Z2y為C3y與C4y之中的最小值，亦即，Z2y=min(C3y,C4y)；Z2y is the minimum value between C3y and C4y, that is, Z2y=min(C3y,C4y);

Z3y為C1y與C4y之中的最小值，亦即，Z3y=min(C1y,C4y)；Z3y is the minimum value between C1y and C4y, that is, Z3y=min(C1y,C4y);

Z4y為C2y與C3y之中的最小值，亦即，Z4y=min(C2y,C3y)；Z4y is the minimum value between C2y and C3y, that is, Z4y=min(C2y,C3y);

Z5y為C1y與C3y之中的最小值，亦即，Z5y=min(C1y,C3y)；Z5y is the minimum value between C1y and C3y, that is, Z5y=min(C1y,C3y);

Z6y為C2y與C4y之中的最小值，亦即，Z6y=min(C2y,C4y)；Z6y is the minimum value between C2y and C4y, that is, Z6y=min(C2y,C4y);

Z1x為C1x與C2x之中的最小值，亦即，Z1x=min(C1x,C2x)；Z1x is the minimum value between C1x and C2x, that is, Z1x=min(C1x,C2x);

Z2x為C3x與C4x之中的最小值，亦即，Z2x=min(C3x,C4x)；Z2x is the minimum value between C3x and C4x, that is, Z2x=min(C3x,C4x);

Z3x為C1x與C4x之中的最小值，亦即，Z3x=min(C1x,C4x)；Z3x is the minimum value between C1x and C4x, that is, Z3x=min(C1x,C4x);

Z4x為C2x與C3x之中的最小值，亦即，Z4x=min(C2x,C3x)；Z4x is the minimum value between C2x and C3x, that is, Z4x=min(C2x,C3x);

Z5x為C1x與C3x之中的最小值，亦即，Z5x=min(C1x,C3x)；Z5x is the minimum value between C1x and C3x, that is, Z5x=min(C1x,C3x);

Z6x為C2x與C4x之中的最小值，亦即，Z6x=min(C2x,C4x)；Z6x is the minimum value between C2x and C4x, that is, Z6x=min(C2x,C4x);

其中：in:

C1y為第一預定象限中的資料數量與該第一預定象限及第二預定象限中的資料數量之總合的比值；C1y is the ratio of the amount of data in the first predetermined quadrant to the sum of the amount of data in the first predetermined quadrant and the second predetermined quadrant;

C2y為第二預定象限中的資料數量與該第一預定象限及第二預定象限中的資料數量之總合的比值；C2y is the ratio of the amount of data in the second predetermined quadrant to the sum of the amounts of data in the first predetermined quadrant and the second predetermined quadrant;

C3y為第三預定象限中的資料數量與該第三預定象限及第四預定象限中的資料數量之總合的比值；C3y is the ratio of the amount of data in the third predetermined quadrant to the sum of the amounts of data in the third predetermined quadrant and the fourth predetermined quadrant;

C4y為第四預定象限中的資料數量與該第三預定象限及第四預定象限中的資料數量之總合的比值；C4y is the ratio of the amount of data in the fourth predetermined quadrant to the sum of the amounts of data in the third predetermined quadrant and the fourth predetermined quadrant;

C1x為第一預定象限中的資料數量與該第一預定象限及第四預定象限中的資料數量之總合的比值；C1x is the ratio of the amount of data in the first predetermined quadrant to the sum of the amount of data in the first predetermined quadrant and the fourth predetermined quadrant;

C2x為第二預定象限中的資料數量與該第二預定象限及第三預定象限中的資料數量之總合的比值；C2x is the ratio of the amount of data in the second predetermined quadrant to the sum of the amount of data in the second predetermined quadrant and the third predetermined quadrant;

C3x為第三預定象限中的資料數量與該第二預定象限及第三預定象限中的資料數量之總合的比值；C3x is the ratio of the amount of data in the third predetermined quadrant to the sum of the amount of data in the second predetermined quadrant and the third predetermined quadrant;

C4x為第四預定象限中的資料數量與該第一預定象限及第四預定象限中的資料數量之總合的比值。C4x is the ratio of the amount of data in the fourth predetermined quadrant to the sum of the amount of data in the first predetermined quadrant and the fourth predetermined quadrant.

為便於說明，簡化為以圖6中的7個搜尋原點O1~O7為示例，表1-1、表1-2與表1-3為基於各搜尋原點O1~O7的數值。各數值的小數之位數大於六位者，以四捨五入取到小數點以下第六位呈現。For the sake of explanation, the seven search origins O1 to O7 in Figure 6 are used as examples. Tables 1-1, 1-2, and 1-3 are the values based on the search origins O1 to O7. If the decimal places of each value are greater than six, they are rounded to the sixth decimal place.

表1-1 基於各搜尋原點的數值 資料數量 搜尋原點 第一 預定象限 第二 預定象限 第三 預定象限 第四 預定象限 C1y C2y C3y C4y O1 7 5 1 1 0.583333 0.416667 0.5 0.5 O2 6 6 1 1 0.5 0.5 0.5 0.5 O3 5 7 1 1 0.416667 0.583333 0.5 0.5 O4 4 8 1 1 0.333333 0.666667 0.5 0.5 O5 5 6 1 2 0.454545 0.545455 0.333333 0.666667 O6 5 5 2 2 0.5 0.5 0.5 0.5 O7 2 4 3 5 0.333333 0.666667 0.375 0.625 Table 1-1 Values based on each search origin Amount of data Search origin First Predetermined Quadrant Second scheduled quadrant The third predetermined quadrant Fourth Predetermined Quadrant C1y C2y C3y C4y O1 7 5 1 1 0.583333 0.416667 0.5 0.5 O2 6 6 1 1 0.5 0.5 0.5 0.5 O3 5 7 1 1 0.416667 0.583333 0.5 0.5 O4 4 8 1 1 0.333333 0.666667 0.5 0.5 O5 5 6 1 2 0.454545 0.545455 0.333333 0.666667 O6 5 5 2 2 0.5 0.5 0.5 0.5 O7 2 4 3 5 0.333333 0.666667 0.375 0.625

表1-2 基於各搜尋原點的數值 資料數量 搜尋原點 第一 預定象限 第二 預定象限 第三 預定象限 第四 預定象限 C1x C2x C3x C4x O1 7 5 1 1 0.875 0.833333 0.166667 0.125 O2 6 6 1 1 0.857143 0.857143 0.142857 0.142857 O3 5 7 1 1 0.833333 0.875 0.125 0.166667 O4 4 8 1 1 0.8 0.888889 0.111111 0.2 O5 5 6 1 2 0.714286 0.857143 0.142857 0.285714 O6 5 5 2 2 0.714286 0.714286 0.285714 0.285714 O7 2 4 3 5 0.285714 0.571429 0.428571 0.714286 Table 1-2 Values based on each search origin Amount of data Search origin First Predetermined Quadrant Second scheduled quadrant The third predetermined quadrant Fourth Predetermined Quadrant C1x C2x C3x C4x O1 7 5 1 1 0.875 0.833333 0.166667 0.125 O2 6 6 1 1 0.857143 0.857143 0.142857 0.142857 O3 5 7 1 1 0.833333 0.875 0.125 0.166667 O4 4 8 1 1 0.8 0.888889 0.111111 0.2 O5 5 6 1 2 0.714286 0.857143 0.142857 0.285714 O6 5 5 2 2 0.714286 0.714286 0.285714 0.285714 O7 2 4 3 5 0.285714 0.571429 0.428571 0.714286

表1-3 基於各搜尋原點的數值 搜尋原點 Z1y Z1x Z2y Z2x Z3y Z3x O1 0.416667 0.833333 0.5 0.125 0.5 0.125 O2 0.5 0.857143 0.5 0.142857 0.5 0.142857 O3 0.416667 0.833333 0.5 0.125 0.416667 0.166667 O4 0.333333 0.8 0.5 0.111111 0.333333 0.2 O5 0.454545 0.714286 0.333333 0.142857 0.454545 0.285714 O6 0.5 0.714286 0.5 0.285714 0.5 0.285714 O7 0.333333 0.285714 0.375 0.428571 0.333333 0.285714 搜尋原點 Z4y Z4x Z5y Z5x Z6y Z6x O1 0.416667 0.166667 0.5 0.166667 0.416667 0.125 O2 0.5 0.142857 0.5 0.142857 0.5 0.142857 O3 0.5 0.125 0.416667 0.125 0.5 0.166667 O4 0.5 0.111111 0.333333 0.111111 0.5 0.2 O5 0.333333 0.142857 0.333333 0.142857 0.545455 0.285714 O6 0.5 0.285714 0.5 0.285714 0.5 0.285714 O7 0.375 0.428571 0.333333 0.285714 0.625 0.571429 Table 1-3 Values based on each search origin Search origin ZlUT Z1x ZlUT Z2x Z3 Z3x O1 0.416667 0.833333 0.5 0.125 0.5 0.125 O2 0.5 0.857143 0.5 0.142857 0.5 0.142857 O3 0.416667 0.833333 0.5 0.125 0.416667 0.166667 O4 0.333333 0.8 0.5 0.111111 0.333333 0.2 O5 0.454545 0.714286 0.333333 0.142857 0.454545 0.285714 O6 0.5 0.714286 0.5 0.285714 0.5 0.285714 O7 0.333333 0.285714 0.375 0.428571 0.333333 0.285714 Search origin Z4 Z4x Z5 Z5x Z6 Z6x O1 0.416667 0.166667 0.5 0.166667 0.416667 0.125 O2 0.5 0.142857 0.5 0.142857 0.5 0.142857 O3 0.5 0.125 0.416667 0.125 0.5 0.166667 O4 0.5 0.111111 0.333333 0.111111 0.5 0.2 O5 0.333333 0.142857 0.333333 0.142857 0.545455 0.285714 O6 0.5 0.285714 0.5 0.285714 0.5 0.285714 O7 0.375 0.428571 0.333333 0.285714 0.625 0.571429

步驟S133：由該些搜尋原點O之Z1y、Z2y、Z3y、Z4y、Z5y、Z6y、Z1x、Z2x、Z3x、Z4x、Z5x與Z6x數值中取最大者的搜尋原點O據以設定該分割原點Os。Step S133: The segmentation origin Os is set based on the search origin O with the largest value among the search origins Z1y, Z2y, Z3y, Z4y, Z5y, Z6y, Z1x, Z2x, Z3x, Z4x, Z5x and Z6x.

如表1-3，處理單元20計算得到每個搜尋原點O1~O7的Z1y、Z2y、Z3y、Z4y、Z5y、Z6y、Z1x、Z2x、Z3x、Z4x、Z5x與Z6x數值後，取當中最大的一者，即Z1x=0.857143，再取其搜尋原點O2可得到預定第一值為0.5、預定第二值為0.3，並且將搜尋原點O2設定為分割原點Os，此分割原點可稱為較佳的分割原點，基於此分割原點所判斷出的資料關係的正確性較高。實務上，可用Argmax函數取得搜尋原點O2的預定第一值與預定第二值。As shown in Table 1-3, the processing unit 20 calculates the values of Z1y, Z2y, Z3y, Z4y, Z5y, Z6y, Z1x, Z2x, Z3x, Z4x, Z5x and Z6x of each search origin O1~O7, and takes the largest one, namely Z1x=0.857143, and then takes the search origin O2 to obtain the predetermined first value of 0.5 and the predetermined second value of 0.3, and sets the search origin O2 as the segmentation origin Os. This segmentation origin can be called a better segmentation origin, and the accuracy of the data relationship determined based on this segmentation origin is higher. In practice, the Argmax function can be used to obtain the predetermined first value and the predetermined second value of the search origin O2.

在一實施例中，該些搜尋原點O之Z1y、Z2y、Z3y、Z4y、Z5y、Z6y、Z1x、Z2x、Z3x、Z4x、Z5x與Z6x數值中取最大者的搜尋原點O為複數個時（例如同時有多個搜尋原點O的Z1y、Z2y、Z3y、Z4y、Z5y、Z6y、Z1x、Z2x、Z3x、Z4x、Z5x與Z6x數值皆相同），則該處理單元20將對應的複數個該搜尋原點O的預定第一值平均及預定第二值平均後分別設定為該分割原點Os的原點第一值與一原點第二值。進一步，如圖7所示，該處理單元20基於對應的複數個該搜尋原點O形成一分界線M，及/或作回歸分析（例如線性回歸、多項式回歸…等），而得到該分界線M的函數。該處理單元20將該分界線M的函數儲存在該儲存單元30及/或直接輸出給通訊單元10，經由通訊介面10a輸出到該外部裝置100。In one embodiment, when there are multiple search origins O whose Z1y, Z2y, Z3y, Z4y, Z5y, Z6y, Z1x, Z2x, Z3x, Z4x, Z5x and Z6x values are the largest (for example, there are multiple search origins O whose Z1y, Z2y, Z3y, Z4y, Z5y, Z6y, Z1x, Z2x, Z3x, Z4x, Z5x and Z6x values are all the same), the processing unit 20 will set the average of the corresponding predetermined first values and the predetermined second values of the multiple search origins O as the origin first value and the origin second value of the segmentation origin Os, respectively. Furthermore, as shown in FIG. 7 , the processing unit 20 forms a dividing line M based on the corresponding plurality of search origins O, and/or performs regression analysis (e.g., linear regression, polynomial regression, etc.), to obtain a function of the dividing line M. The processing unit 20 stores the function of the dividing line M in the storage unit 30 and/or directly outputs it to the communication unit 10, and outputs it to the external device 100 via the communication interface 10a.

步驟S14：資料分析裝置1依據該分割原點Os的該第一分割線L1與該第二分割線L2將該些資料區分為四個象限，該些象限中分別具有該些資料的一部分。Step S14: The data analysis device 1 divides the data into four quadrants according to the first dividing line L1 and the second dividing line L2 of the dividing origin Os, and each of the quadrants contains a portion of the data.

復如圖3所示，該處理單元20基於該分割原點Os所區分的該四個象限分別為一第一象限Q1、一第二象限Q2、一第三象限Q3與一第四象限Q4。As shown in FIG. 3 again, the four quadrants divided by the processing unit 20 based on the segmentation origin Os are a first quadrant Q1, a second quadrant Q2, a third quadrant Q3 and a fourth quadrant Q4.

步驟S15：資料分析裝置1依據基於該分割原點Os的各該象限中的資料數量與相鄰的象限中的資料數量的比值判斷該些資料的一資料關係，並輸出基於該分割原點Os的該資料關係。Step S15: The data analysis device 1 determines a data relationship of the data based on the ratio of the data quantity in each quadrant based on the segmentation origin Os to the data quantity in the adjacent quadrant, and outputs the data relationship based on the segmentation origin Os.

本實施例中，係由該處理單元20判斷該資料關係，該處理單元20輸出該資料關係至該通訊單元10，該通訊單元10將該資料關係輸出至該通訊介面10a，藉以傳輸到外部裝置100。In this embodiment, the data relationship is determined by the processing unit 20, and the processing unit 20 outputs the data relationship to the communication unit 10. The communication unit 10 outputs the data relationship to the communication interface 10a to be transmitted to the external device 100.

該資料關係為複數個資料關係中的一者，該些資料關係包括水平關係、垂直關係、正比關係與反比關係，於後分別說明。The data relationship is one of a plurality of data relationships, including a horizontal relationship, a vertical relationship, a proportional relationship, and an inverse proportional relationship, which will be described below.

〔水平關係〕〔Horizontal Relationship〕

水平關係有兩種，第一種是多數資料落於第一象限Q1與第二象限Q2，第三象限Q3與第四象限Q4的資料相對較少；第二種是多數資料落於第三象限Q3與第四象限Q4，第一象限Q1與第二象限Q2的資料相對較少。There are two types of horizontal relationships. The first is that most of the data falls in the first quadrant Q1 and the second quadrant Q2, and the data in the third quadrant Q3 and the fourth quadrant Q4 are relatively small. The second is that most of the data falls in the third quadrant Q3 and the fourth quadrant Q4, and the data in the first quadrant Q1 and the second quadrant Q2 are relatively small.

〔第一種水平關係〕〔The first horizontal relationship〕

該處理單元20係於符合下列條件時判斷該資料關係為第一種水平關係：The processing unit 20 determines that the data relationship is a first level relationship when the following conditions are met:

C1≥P1 且 C2≥P2；C1≥P1 and C2≥P2;

其中，C1為第一象限Q1中的資料數量與第一象限Q1及第四象限Q4中的資料數量之總合的比值，P1為一第一預定比值，0.5≤P1≤0.7；C2為第二象限Q2中的資料數量與第二象限Q2及第三象限Q3中的資料數量之總合的比值，P2為一第二預定比值，0.5≤P2≤0.7。本實施例中P1以0.5（50%）為例，P2以0.5（50%）為例，符合上述條件代表第一象限Q1及第四象限Q4中有50%以上的資料是在第一象限Q1，第二象限Q2及第三象限Q3中有50%以上的資料是在第二象限Q2。Wherein, C1 is the ratio of the amount of data in the first quadrant Q1 to the total amount of data in the first quadrant Q1 and the fourth quadrant Q4, P1 is a first predetermined ratio, 0.5≤P1≤0.7; C2 is the ratio of the amount of data in the second quadrant Q2 to the total amount of data in the second quadrant Q2 and the third quadrant Q3, P2 is a second predetermined ratio, 0.5≤P2≤0.7. In this embodiment, P1 is 0.5 (50%) as an example, and P2 is 0.5 (50%) as an example. Meeting the above conditions means that more than 50% of the data in the first quadrant Q1 and the fourth quadrant Q4 are in the first quadrant Q1, and more than 50% of the data in the second quadrant Q2 and the third quadrant Q3 are in the second quadrant Q2.

以圖3為例，請參下表2。C1大於P1，C2大於P2，因此，處理單元20判斷為第一種水平關係。處理單元20更計算C3與C4，其中，C3為第三象限Q3中的資料數量與第二象限Q2及第三象限Q3中的資料數量之總合的比值，C4為第四象限Q4中的資料數量與第一象限Q1及第四象限Q4中的資料數量之總合的比值。更進一步，處理單元20可再判斷C1大於C4且C2大於C3，或C4小於P1且C3小於P2，以進一步確定該資料關係為第一種水平關係。Take FIG. 3 as an example, please refer to Table 2 below. C1 is greater than P1, and C2 is greater than P2, so the processing unit 20 determines that it is a first horizontal relationship. The processing unit 20 further calculates C3 and C4, where C3 is the ratio of the amount of data in the third quadrant Q3 to the sum of the amount of data in the second quadrant Q2 and the third quadrant Q3, and C4 is the ratio of the amount of data in the fourth quadrant Q4 to the sum of the amount of data in the first quadrant Q1 and the fourth quadrant Q4. Furthermore, the processing unit 20 can further determine that C1 is greater than C4 and C2 is greater than C3, or C4 is less than P1 and C3 is less than P2, to further determine that the data relationship is a first horizontal relationship.

表2 第一種水平關係的數值示例 資料數量 第一象限 第二象限 第三象限 第四象限 C1 C2 C3 C4 6 6 1 1 0.857143 0.857143 0.142857 0.142857 Table 2 Examples of numerical values for the first level of relationship Amount of data Quadrant 1 Second Quadrant The third quadrant Quadrant 4 C1 C2 C3 C4 6 6 1 1 0.857143 0.857143 0.142857 0.142857

該處理單元20輸出第一種水平關係至該通訊單元10，該通訊單元10將水平關係輸出至該通訊介面10a，藉以傳輸到外部裝置100。The processing unit 20 outputs the first horizontal relationship to the communication unit 10 , and the communication unit 10 outputs the horizontal relationship to the communication interface 10 a to be transmitted to the external device 100 .

處理單元20確定資料關係為水平關係後，更可產生一水平分割線Lh，請配合圖8，說明產生該水平分割線Lh的步驟，其包含：After the processing unit 20 determines that the data relationship is a horizontal relationship, it can further generate a horizontal dividing line Lh. Referring to FIG. 8 , the steps of generating the horizontal dividing line Lh include:

選定該分割原點Os的原點第一值，並且由該下限第二值與該上限第二值之間選定與該原點第一值配合的複數個參考第二值，以形成複數個參考原點O2、O5、O6、O7（本實施例中參考原點為選自搜尋原點O2、O5、O6、O7），各該參考原點O2、O5、O6、O7具有一參考第一值，各該參考第一值為該原點第一值。分割原點Os也構成其中一個參考原點。The first origin value of the segmentation origin Os is selected, and a plurality of reference second values matching the first origin value are selected between the lower limit second value and the upper limit second value to form a plurality of reference origins O2, O5, O6, O7 (in this embodiment, the reference origin is selected from the search origins O2, O5, O6, O7), each of which has a reference first value, and each of which is the first origin value. The segmentation origin Os also constitutes one of the reference origins.

以各該參考原點O2、O5、O6、O7的參考第一值與參考第二值分別形成一第一參考分割線Lr1與一第二參考分割線Lr2，且各該參考原點的第一參考分割線Lr1與第二參考分割線Lr2將該些資料區分為四個參考象限，以圖8參考原點O6為例。該四個參考象限分別為一第一參考象限Qr1、一第二參考象限Qr2、一第三參考象限Qr3與一第四參考象限Qr4。The reference first value and the reference second value of each reference origin O2, O5, O6, O7 form a first reference dividing line Lr1 and a second reference dividing line Lr2, respectively, and the first reference dividing line Lr1 and the second reference dividing line Lr2 of each reference origin divide the data into four reference quadrants, taking the reference origin O6 in FIG8 as an example. The four reference quadrants are a first reference quadrant Qr1, a second reference quadrant Qr2, a third reference quadrant Qr3, and a fourth reference quadrant Qr4.

分別計算基於各該參考原點O2、O5、O6、O7的Csum1數值，所使用的算式如式(5)所示。較佳者，以四個參考象限皆有資料的參考原點O2、O5、O6、O7計算Csum1數值。The Csum1 values based on the reference origins O2, O5, O6, and O7 are calculated respectively, and the formula used is shown in formula (5). Preferably, the Csum1 values are calculated using the reference origins O2, O5, O6, and O7, which have data in all four reference quadrants.

Csum1=C11x+C21x  ……. (5)Csum1=C11x+C21x  ……. (5)

其中：in:

C11x為第一參考象限Qr1中的資料數量與第一參考象限Qr1及第四參考象限Qr4中的資料數量之總合的比值；C11x is the ratio of the amount of data in the first reference quadrant Qr1 to the sum of the amount of data in the first reference quadrant Qr1 and the fourth reference quadrant Qr4;

C21x為第二參考象限Qr2中的資料數量與第二參考象限Qr2及第三參考象限Qr3中的資料數量之總合的比值。C21x is the ratio of the amount of data in the second reference quadrant Qr2 to the total amount of data in the second reference quadrant Qr2 and the third reference quadrant Qr3.

為便於說明，以圖8中的4個參考原點O2、O5、O6、O7為示例，表3為基於各參考原點O2、O5、O6、O7的數值。For ease of explanation, taking the four reference origins O2, O5, O6, and O7 in FIG. 8 as an example, Table 3 shows the values based on the reference origins O2, O5, O6, and O7.

表3 基於各參考原點的數值 資料數量 參考原點 第一 參考象限 第二 參考象限 第三 參考象限 第四 參考象限 C11x C21x Csum1 差值 O2 6 6 1 1 0.857143 0.857143 1.714286 0.142857 O5 5 6 1 2 0.714286 0.857143 1.571429 0.142857 O6 5 5 2 2 0.714286 0.714286 1.428571 0.571429 O7 2 4 3 5 0.285714 0.571429 0.857143 - Table 3 Values based on various reference points Amount of data Reference origin First reference quadrant Second reference quadrant Third reference quadrant Fourth reference quadrant C11x C21x Csum1 Difference O2 6 6 1 1 0.857143 0.857143 1.714286 0.142857 O5 5 6 1 2 0.714286 0.857143 1.571429 0.142857 O6 5 5 2 2 0.714286 0.714286 1.428571 0.571429 O7 2 4 3 5 0.285714 0.571429 0.857143 -

計算各該參考原點O2、O5、O6、O7的Csum1與其相鄰且參考第二值較大之參考原點的Csum1之差值，依據具有最大之差值的參考原點的參考第二值設定一水平分割線。The difference between the Csum1 of each reference origin O2, O5, O6, O7 and the Csum1 of the adjacent reference origin with a larger reference second value is calculated, and a horizontal dividing line is set according to the reference second value of the reference origin with the largest difference.

如表3中所示，具最大之差值（0.571429）的是參考原點O6，因此，依據參考原點O6的參考第二值（0.5）設定為水平分割線Lh，並可將水平分割線Lh的值輸出到該通訊單元10，再由通訊單元10輸出到通訊介面10a。最大之差值代表資料在參考原點O6以上的資料與以下的資料數量變化有明顯的差異，因此，設定水平分割線Lh後，可更區分出資料數量的變化。處理單元20可更進一步以水平分割線Lh將該些資料分為兩類。As shown in Table 3, the reference origin O6 has the largest difference (0.571429). Therefore, the reference second value (0.5) of the reference origin O6 is set as the horizontal dividing line Lh, and the value of the horizontal dividing line Lh can be output to the communication unit 10, and then output to the communication interface 10a by the communication unit 10. The largest difference represents that the data above the reference origin O6 and the data below have a significant difference in the amount of change. Therefore, after setting the horizontal dividing line Lh, the change in the amount of data can be further distinguished. The processing unit 20 can further divide the data into two categories using the horizontal dividing line Lh.

〔第二種水平關係〕[Second Horizontal Relationship]

該處理單元20係於符合下列條件時判斷該資料關係為第二種水平關係：The processing unit 20 determines that the data relationship is a second level relationship when the following conditions are met:

C4≥P1 且 C3≥P2；C4≥P1 and C3≥P2;

以圖9及表4的情況為例，在分割原點Os處，C4大於P1，C3大於P2，因此，處理單元20判斷為第二種水平關係。更進一步，處理單元20可再判斷C4大於C1且C3大於C2，或C1小於P1且C2小於P2，以進一步確定是第二種水平關係。Taking the situation of FIG. 9 and Table 4 as an example, at the segmentation origin Os, C4 is greater than P1, and C3 is greater than P2, so the processing unit 20 determines that it is the second horizontal relationship. Furthermore, the processing unit 20 can further determine that C4 is greater than C1 and C3 is greater than C2, or C1 is less than P1 and C2 is less than P2, to further determine that it is the second horizontal relationship.

表4 第二種水平關係的數值示例 資料數量 第一象限 第二象限 第三象限 第四象限 C1 C2 C3 C4 1 1 6 6 0.142857 0.142857 0.857143 0.857143 Table 4 Examples of numerical values for the second level relationship Amount of data Quadrant 1 Second Quadrant The third quadrant Quadrant 4 C1 C2 C3 C4 1 1 6 6 0.142857 0.142857 0.857143 0.857143

該處理單元20輸出第二種水平關係至該通訊單元10，該通訊單元10將水平關係輸出至該通訊介面10a，藉以傳輸到外部裝置100。The processing unit 20 outputs the second horizontal relationship to the communication unit 10 , and the communication unit 10 outputs the horizontal relationship to the communication interface 10 a to be transmitted to the external device 100 .

處理單元20確定資料關係為水平關係後，更可產生一水平分割線Lh。請配合圖9，說明產生該水平分割線Lh的步驟，其步驟與第一種水平關係的水平分割線Lh的產生步驟大致相同，差異在於Csum2的計算方式不同，以及計算差值的方式不同，亦即：After the processing unit 20 determines that the data relationship is a horizontal relationship, a horizontal dividing line Lh can be generated. Please refer to FIG. 9 to explain the steps of generating the horizontal dividing line Lh. The steps are roughly the same as the steps of generating the horizontal dividing line Lh of the first horizontal relationship. The difference lies in the different calculation methods of Csum2 and the different methods of calculating the difference, that is:

Csum2=C31x+C41xCsum2=C31x+C41x

其中：in:

C31x為第三參考象限Qr3中的資料數量與第二參考象限Qr2及第三參考象限Qr3中的資料數量之總合的比值；C31x is the ratio of the amount of data in the third reference quadrant Qr3 to the sum of the amount of data in the second reference quadrant Qr2 and the third reference quadrant Qr3;

C41x為第四參考象限Qr4中的資料數量與第一參考象限Qr1及第四參考象限Qr4中的資料數量之總合的比值。C41x is the ratio of the amount of data in the fourth reference quadrant Qr4 to the total amount of data in the first reference quadrant Qr1 and the fourth reference quadrant Qr4.

各參考原點O2,O5,O6,O7的數值示例於表5。Examples of the values of the reference points O2, O5, O6, and O7 are shown in Table 5.

表5 基於各參考原點的數值 資料數量 參考原點 第一 參考象限 第二 參考象限 第三 參考象限 第四 參考象限 C31x C41x Csum2 差值 O7 5 3 4 2 0.571429 0.285714 0.857143 - O6 2 2 5 5 0.714286 0.714286 1.428571 0.571429 O5 2 1 6 5 0.857143 0.714286 1.571429 0.142857 O2 1 1 6 6 0.857143 0.857143 1.714286 0.142857 Table 5 Values based on various reference points Amount of data Reference origin First reference quadrant Second reference quadrant Third reference quadrant Fourth reference quadrant C31x C41x Csum2 Difference O7 5 3 4 2 0.571429 0.285714 0.857143 - O6 2 2 5 5 0.714286 0.714286 1.428571 0.571429 O5 2 1 6 5 0.857143 0.714286 1.571429 0.142857 O2 1 1 6 6 0.857143 0.857143 1.714286 0.142857

計算各該參考原點O2,O5,O6,O7的Csum2與其相鄰且參考第二值較小之參考原點的Csum2之差值，依據具有最大之差值的參考原點的參考第二值設定一水平分割線Lh。The difference between the Csum2 of each reference origin O2, O5, O6, O7 and the Csum2 of the adjacent reference origin with a smaller reference second value is calculated, and a horizontal dividing line Lh is set according to the reference second value of the reference origin with the largest difference.

如表5中所示，具有最大之差值（0.571429）的為參考原點O6，因此，依據參考原點O6的參考第二值（0.5）設定為水平分割線Lh，並可將水平分割線Lh的值輸出到該通訊單元10，再由通訊單元10輸出到通訊介面10a。最大之差值代表資料在參考原點O6以上的資料與以下的資料數量變化有明顯的差異，因此，設定水平分割線Lh後，可更區分出資料數量的變化。處理單元20可更進一步以水平分割線Lh將該些資料分為兩類。As shown in Table 5, the reference origin O6 has the largest difference (0.571429). Therefore, the horizontal dividing line Lh is set according to the reference second value (0.5) of the reference origin O6, and the value of the horizontal dividing line Lh can be output to the communication unit 10, and then output to the communication interface 10a by the communication unit 10. The largest difference represents that the data above the reference origin O6 and the data below have a significant difference in the amount of change. Therefore, after setting the horizontal dividing line Lh, the change in the amount of data can be further distinguished. The processing unit 20 can further divide the data into two categories using the horizontal dividing line Lh.

〔垂直關係〕〔Vertical relationship〕

垂直關係有兩種，第一種是多數資料落於第一象限Q1與第四象限Q4，第二象限Q2與第三象限Q3的資料相對較少；第二種是多數資料落於第二象限Q2與第三象限Q3，第一象限Q1與第四象限Q4的資料相對較少。There are two types of vertical relationships. The first is that most of the data falls in the first quadrant Q1 and the fourth quadrant Q4, and the data in the second quadrant Q2 and the third quadrant Q3 are relatively small. The second is that most of the data falls in the second quadrant Q2 and the third quadrant Q3, and the data in the first quadrant Q1 and the fourth quadrant Q4 are relatively small.

〔第一種垂直關係〕〔The first vertical relationship〕

請參圖10，該處理單元20係於符合下列條件時判斷該資料關係為第一種垂直關係：Referring to FIG. 10 , the processing unit 20 determines that the data relationship is a first vertical relationship when the following conditions are met:

C1’≥P1 且 C4’≥P2；C1’≥P1 and C4’≥P2;

其中，C1’為第一象限Q1中的資料數量與第一象限Q1及第二象限Q2中的資料數量之總合的比值，P1為一第一預定比值，0.5≤P1≤0.7；C4’為第四象限Q4中的資料數量與第三象限Q3及第四象限Q4中的資料數量之總合的比值，P2為一第二預定比值，0.5≤P2≤0.7。本實施例中P1以0.5（50%）為例，P2以0.5（50%）為例，符合上述條件代表第一象限Q1及第二象限Q2中有50%以上的資料是在第一象限Q1，第三象限Q3及第四象限Q4中有50%以上的資料是在第四象限Q4。Among them, C1' is the ratio of the amount of data in the first quadrant Q1 to the total amount of data in the first quadrant Q1 and the second quadrant Q2, P1 is a first predetermined ratio, 0.5≤P1≤0.7; C4' is the ratio of the amount of data in the fourth quadrant Q4 to the total amount of data in the third quadrant Q3 and the fourth quadrant Q4, P2 is a second predetermined ratio, 0.5≤P2≤0.7. In this embodiment, P1 is 0.5 (50%) as an example, and P2 is 0.5 (50%) as an example. Meeting the above conditions means that more than 50% of the data in the first quadrant Q1 and the second quadrant Q2 are in the first quadrant Q1, and more than 50% of the data in the third quadrant Q3 and the fourth quadrant Q4 are in the fourth quadrant Q4.

請參圖10，在分割原點Os處，請參下表6，C1’大於P1，C4’大於P2，因此，處理單元20判斷為第一種垂直關係。Please refer to FIG. 10 . At the segmentation origin Os, please refer to Table 6 below. C1′ is greater than P1, and C4′ is greater than P2. Therefore, the processing unit 20 determines that it is a first vertical relationship.

處理單元20計算C2’與C3’，其中，C2’為第二象限Q2中的資料數量與第一象限Q1及第二象限Q2中的資料數量之總合的比值，C3’為第三象限Q3中的資料數量與第三象限Q3及第四象限Q4中的資料數量之總合的比值。更進一步，處理單元20可再判斷C1’大於C2’且C3’大於C4’，或C2’小於P1且C3’小於P2，以進一步確定該資料關係為第一種垂直關係。The processing unit 20 calculates C2' and C3', wherein C2' is the ratio of the amount of data in the second quadrant Q2 to the sum of the amount of data in the first quadrant Q1 and the second quadrant Q2, and C3' is the ratio of the amount of data in the third quadrant Q3 to the sum of the amount of data in the third quadrant Q3 and the fourth quadrant Q4. Furthermore, the processing unit 20 may further determine that C1' is greater than C2' and C3' is greater than C4', or that C2' is less than P1 and C3' is less than P2, to further determine that the data relationship is the first vertical relationship.

表6 第一種垂直關係的數值示例 資料數量 第一象限 第二象限 第三象限 第四象限 C1’ C2’ C3’ C4’ 6 1 1 6 0.857143 0.142857 0.142857 0.857143 Table 6 Examples of numerical values for the first vertical relationship Amount of data Quadrant 1 Second Quadrant The third quadrant Quadrant 4 C1' C2' C3' C4' 6 1 1 6 0.857143 0.142857 0.142857 0.857143

該處理單元20輸出第一種垂直關係至該通訊單元10，該通訊單元10將垂直關係輸出至該通訊介面10a，藉以傳輸到外部裝置100。The processing unit 20 outputs the first vertical relationship to the communication unit 10 , and the communication unit 10 outputs the vertical relationship to the communication interface 10 a to be transmitted to the external device 100 .

處理單元20確定資料關係為垂直關係後，更可產生一垂直分割線Lv，請配合圖11，說明產生該垂直分割線Lv的步驟，其包含：After the processing unit 20 determines that the data relationship is a vertical relationship, it can further generate a vertical dividing line Lv. Referring to FIG. 11 , the steps of generating the vertical dividing line Lv are described, which include:

選定該分割原點Os的原點第二值，並且由該下限第一值與該上限第一值之間選定與該原點第二值配合的複數個參考第一值，以形成複數個參考原點O2,O5,O6,O7，各該參考原點O2,O5,O6,O7具有一參考第二值，各該參考第二值為該原點第二值。分割原點Os也構成其中一個參考原點O2。The second origin value of the split origin Os is selected, and a plurality of reference first values matching the second origin value are selected between the lower limit first value and the upper limit first value to form a plurality of reference origins O2, O5, O6, O7, each of which has a reference second value, and each of which is the second origin value. The split origin Os also constitutes one of the reference origins O2.

以各該參考原點O2,O5,O6,O7的參考第一值與參考第二值分別形成一第一參考分割線Lr1與一第二參考分割線Lr2，且各該參考原點O2,O5,O6,O7的第一參考分割線Lr1與第二參考分割線Lr2將該些資料區分為第一至第四參考象限Qr1~Qr4；A first reference dividing line Lr1 and a second reference dividing line Lr2 are formed by using the first reference value and the second reference value of each reference origin O2, O5, O6, O7, respectively, and the first reference dividing line Lr1 and the second reference dividing line Lr2 of each reference origin O2, O5, O6, O7 divide the data into first to fourth reference quadrants Qr1-Qr4;

分別計算基於各該參考原點O2,O5,O6,O7的Csum1’數值，所使用的算式如式(6)所示。較佳者，以四個參考象限皆有資料的參考原點O2,O5,O6,O7計算Csum1’數值。The Csum1' values based on the reference origins O2, O5, O6, and O7 are calculated respectively, and the formula used is shown in formula (6). Preferably, the Csum1' values are calculated based on the reference origins O2, O5, O6, and O7 with data in all four reference quadrants.

Csum1’=C11y+C41y  ……. (6)Csum1’=C11y+C41y  ……. (6)

其中，C11y為第一參考象限Qr1中的資料數量與第一參考象限Qr1及第二參考象限Qr2中的資料數量之總合的比值；C41y為第四參考象限Qr4中的資料數量與第三參考象限Qr3及第四參考象限Qr4中的資料數量之總合的比值。Among them, C11y is the ratio of the amount of data in the first reference quadrant Qr1 to the total amount of data in the first reference quadrant Qr1 and the second reference quadrant Qr2; C41y is the ratio of the amount of data in the fourth reference quadrant Qr4 to the total amount of data in the third reference quadrant Qr3 and the fourth reference quadrant Qr4.

為便於說明，以圖11中的4個參考原點O2,O5,O6,O7為示例，表7為基於各參考原點O2,O5,O6,O7的數值。For ease of explanation, taking the four reference origins O2, O5, O6, and O7 in FIG. 11 as an example, Table 7 shows the values based on the reference origins O2, O5, O6, and O7.

表7 基於各參考原點的數值 資料數量 參考原點 第一 參考象限 第二 參考象限 第三 參考象限 第四 參考象限 C11y C41y Csum1’ 差值 O2 6 1 1 6 0.857143 0.857143 1.714286 0.142857 O5 6 1 2 5 0.857143 0.714286 1.571429 0.142857 O6 5 2 2 5 0.714286 0.714286 1.428571 0.571429 O7 4 3 5 2 0.571429 0.285714 0.857143 - Table 7 Values based on various reference points Amount of data Reference origin First reference quadrant Second reference quadrant Third reference quadrant Fourth reference quadrant C11y C41y Csum1' Difference O2 6 1 1 6 0.857143 0.857143 1.714286 0.142857 O5 6 1 2 5 0.857143 0.714286 1.571429 0.142857 O6 5 2 2 5 0.714286 0.714286 1.428571 0.571429 O7 4 3 5 2 0.571429 0.285714 0.857143 -

計算各該參考原點的Csum1’與其相鄰且參考第一值較大之參考原點的Csum1’之差值，依據具有最大之差值的參考原點的參考第一值設定一垂直分割線Lv。The difference between the Csum1' of each reference origin and the Csum1' of the adjacent reference origin with a larger reference first value is calculated, and a vertical dividing line Lv is set according to the reference first value of the reference origin with the largest difference.

如表7中所示，具有最大之差值（0.571429）的是參考原點O6，因此，依據參考原點O6的參考第二值（0.5）設定為垂直分割線Lv，並可將垂直分割線Lv的值輸出到該通訊單元10，再由通訊單元10輸出到通訊介面10a。最大之差值代表資料在參考原點O6以上的資料與以下的資料數量變化有明顯的差異，因此，設定垂直分割線Lv後，可更區分出資料數量的變化。處理單元可更進一步以垂直分割線Lv將該些資料分為兩類。As shown in Table 7, the reference origin O6 has the largest difference (0.571429). Therefore, the reference second value (0.5) of the reference origin O6 is set as the vertical dividing line Lv, and the value of the vertical dividing line Lv can be output to the communication unit 10, and then output to the communication interface 10a by the communication unit 10. The largest difference represents that the data above the reference origin O6 and the data below have a significant difference in the amount of change. Therefore, after setting the vertical dividing line Lv, the change in the amount of data can be further distinguished. The processing unit can further divide the data into two categories using the vertical dividing line Lv.

〔第二種垂直關係〕[The second vertical relationship]

該處理單元20係於符合下列條件時判斷該資料關係為第二種垂直關係：The processing unit 20 determines that the data relationship is a second vertical relationship when the following conditions are met:

C2’≥P1 且 C3’≥P2C2’≥P1 and C3’≥P2

以圖12及表8的情況為例，在分割原點處，C2’大於P1，C3’大於P2，因此，處理單元20判斷為第二種垂直關係。更進一步，處理單元20可再判斷C2’大於C1’且C3’大於C4’，或C1’小於P1且C4’小於P2，以進一步確定是第二種垂直關係。Taking the situation of FIG. 12 and Table 8 as an example, at the segmentation origin, C2' is greater than P1, and C3' is greater than P2, so the processing unit 20 determines that it is the second vertical relationship. Furthermore, the processing unit 20 can further determine that C2' is greater than C1' and C3' is greater than C4', or C1' is less than P1 and C4' is less than P2, to further determine that it is the second vertical relationship.

表8 第二種垂直關係的數值示例 資料數量 第一象限 第二象限 第三象限 第四象限 C1’ C2’ C3’ C4’ 1 6 6 1 0.142857 0.857143 0.857143 0.142857 Table 8 Examples of numerical values for the second vertical relationship Amount of data Quadrant 1 Second Quadrant The third quadrant Quadrant 4 C1' C2' C3' C4' 1 6 6 1 0.142857 0.857143 0.857143 0.142857

該處理單元20輸出第二種垂直關係至該通訊單元10，該通訊單元10將垂直關係輸出至該通訊介面10a，藉以傳輸到外部裝置100。The processing unit 20 outputs the second vertical relationship to the communication unit 10 , and the communication unit 10 outputs the vertical relationship to the communication interface 10 a to be transmitted to the external device 100 .

處理單元20確定資料關係為垂直關係後，更可產生一垂直分割線Lv。請配合圖13，說明產生該垂直分割線Lv的步驟，其步驟與第一種垂直關係的垂直分割線的產生步驟大致相同，差異在於Csum2’的計算方式不同，以及計算差值的方式不同，亦即：After the processing unit 20 determines that the data relationship is a vertical relationship, a vertical dividing line Lv can be generated. Please refer to FIG. 13 to explain the steps of generating the vertical dividing line Lv. The steps are roughly the same as the steps of generating the vertical dividing line of the first vertical relationship. The difference lies in the different calculation methods of Csum2' and the different methods of calculating the difference, that is:

Csum2’=C21y+C31yCsum2’=C21y+C31y

其中：in:

C21y為第二參考象限Qr2中的資料數量與第一參考象限Qr1及第二參考象限Qr2中的資料數量之總合的比值；C21y is the ratio of the amount of data in the second reference quadrant Qr2 to the sum of the amount of data in the first reference quadrant Qr1 and the second reference quadrant Qr2;

C31y為第三參考象限Qr3中的資料數量與第三參考象限Qr3及第四參考象限Qr4中的資料數量之總合的比值。C31y is the ratio of the amount of data in the third reference quadrant Qr3 to the total amount of data in the third reference quadrant Qr3 and the fourth reference quadrant Qr4.

各參考原點O2,O5,O6,O7的數值示例於表9。Examples of the values of the reference points O2, O5, O6, and O7 are shown in Table 9.

表9 基於各參考原點的數值 資料數量 參考原點 第一 參考象限 第二 參考象限 第三 參考象限 第四 參考象限 C21y C31y Csum2’ 差值 O7 3 4 2 5 0.571429 0.285714 0.857143 - O6 2 5 5 2 0.714286 0.714286 1.428571 0.571429 O5 1 6 5 2 0.857143 0.714286 1.571429 0.142857 O2 1 6 6 1 0.857143 0.857143 1.714286 0.142857 Table 9 Values based on various reference points Amount of data Reference origin First reference quadrant Second reference quadrant Third reference quadrant Fourth reference quadrant C21y C31y Csum2' Difference O7 3 4 2 5 0.571429 0.285714 0.857143 - O6 2 5 5 2 0.714286 0.714286 1.428571 0.571429 O5 1 6 5 2 0.857143 0.714286 1.571429 0.142857 O2 1 6 6 1 0.857143 0.857143 1.714286 0.142857

計算各該參考原點O2,O5,O6,O7的Csum2’與其相鄰且參考第二值較小之參考原點的Csum2’之差值，依據具有最大之差值的參考原點的參考第二值設定一垂直分割線Lv。The difference between the Csum2' of each reference origin O2, O5, O6, O7 and the Csum2' of the adjacent reference origin with a smaller reference second value is calculated, and a vertical dividing line Lv is set according to the reference second value of the reference origin with the largest difference.

如表9中所示，具有最大之差值（0.571429）的是參考原點O6，因此，依據參考原點O6的參考第二值（0.5）設定為垂直分割線Lv，並可將垂直分割線Lv的值輸出到該通訊單元10，再由通訊單元10輸出到通訊介面10a。最大之差值代表資料在參考原點O6以上的資料與以下的資料數量變化有明顯的差異，因此，設定垂直分割線Lv後，可更區分出資料數量的變化。處理單元20可更進一步以垂直分割線Lv將該些資料分為兩類。As shown in Table 9, the reference origin O6 has the largest difference (0.571429). Therefore, the reference second value (0.5) of the reference origin O6 is set as the vertical dividing line Lv, and the value of the vertical dividing line Lv can be output to the communication unit 10, and then output to the communication interface 10a by the communication unit 10. The largest difference represents that the data above the reference origin O6 and the data below have a significant difference in the amount of change. Therefore, after setting the vertical dividing line Lv, the change in the amount of data can be further distinguished. The processing unit 20 can further divide the data into two categories using the vertical dividing line Lv.

〔正比關係〕〔Proportional relationship〕

正比關係則是多數資料落於第一象限Q1與第三象限Q3，第二象限Q2與第四象限Q4的資料相對較少，圖14參照。The proportional relationship means that most of the data falls in the first quadrant Q1 and the third quadrant Q3, while the data in the second quadrant Q2 and the fourth quadrant Q4 are relatively small, see Figure 14.

該處理單元20係於符合下列條件時判斷該資料關係為正比關係：The processing unit 20 determines that the data relationship is a proportional relationship when the following conditions are met:

C1≥P1 且 C3≥P2；C1≥P1 and C3≥P2;

其中：in:

C1為第一象限Q1中的資料數量與第一象限Q1及其相鄰之第二象限Q2與第四象限Q4之其中一者的資料數量之總合的比值，P1為一第一預定比值，0.5≤P1≤0.7；C1 is the ratio of the amount of data in the first quadrant Q1 to the total amount of data in the first quadrant Q1 and one of the adjacent second quadrants Q2 and fourth quadrants Q4, P1 is a first predetermined ratio, 0.5≤P1≤0.7;

C3為第三象限Q3中的資料數量與第三象限Q3及其相鄰之第二象限Q2與第四象限Q4之其中另一者的資料數量之總合的比值，P2為一第二預定比值，0.5≤P2≤0.7。C3 is the ratio of the amount of data in the third quadrant Q3 to the total amount of data in the third quadrant Q3 and the other of the adjacent second quadrant Q2 and fourth quadrant Q4, and P2 is a second predetermined ratio, 0.5≤P2≤0.7.

本實施例中P1以0.5（50%）為例，P2以0.5（50%）為例，符合上述條件代表第一象限Q1及第四象限Q4中有50%以上的資料是在第一象限Q1，第二象限Q2及第三象限Q3中有50%以上的資料是在第三象限Q3；或者第一象限Q1及第二象限Q2中有50%以上的資料是在第一象限Q1，第三象限Q3及第四象限Q4中有50%以上的資料是在第三象限Q3。In this embodiment, P1 is taken as 0.5 (50%), and P2 is taken as 0.5 (50%). Meeting the above conditions means that more than 50% of the data in the first quadrant Q1 and the fourth quadrant Q4 are in the first quadrant Q1, and more than 50% of the data in the second quadrant Q2 and the third quadrant Q3 are in the third quadrant Q3; or more than 50% of the data in the first quadrant Q1 and the second quadrant Q2 are in the first quadrant Q1, and more than 50% of the data in the third quadrant Q3 and the fourth quadrant Q4 are in the third quadrant Q3.

假設為表10的情況，在分割原點Os處，C1為第一象限Q1中的資料數量與第一象限Q1及第四象限Q4的資料數量之總合的比值，C3為第三象限Q3中的資料數量與第二象限Q2及第三象限Q3的資料數量之總合的比值。C2為第二象限Q2中的資料數量與第二象限Q2及第三象限Q3的資料數量之總合的比值，C4為第四象限Q4中的資料數量與第一象限Q1及第四象限Q4的資料數量之總合的比值。Assuming the situation in Table 10, at the dividing origin Os, C1 is the ratio of the number of data in the first quadrant Q1 to the total number of data in the first quadrant Q1 and the fourth quadrant Q4, C3 is the ratio of the number of data in the third quadrant Q3 to the total number of data in the second quadrant Q2 and the third quadrant Q3. C2 is the ratio of the number of data in the second quadrant Q2 to the total number of data in the second quadrant Q2 and the third quadrant Q3, and C4 is the ratio of the number of data in the fourth quadrant Q4 to the total number of data in the first quadrant Q1 and the fourth quadrant Q4.

C1大於P1，C3大於P2，因此，處理單元20判斷為正比關係。更進一步，處理單元20可再判斷C1大於C4且C3大於C2，或C4小於P1且C2小於P2，以進一步確定是正比關係。C1 is greater than P1, and C3 is greater than P2, so the processing unit 20 determines that it is a proportional relationship. Furthermore, the processing unit 20 can further determine that C1 is greater than C4 and C3 is greater than C2, or C4 is less than P1 and C2 is less than P2, to further determine that it is a proportional relationship.

表10 正比關係的數值示例 資料數量 第一象限 第二象限 第三象限 第四象限 C1 C2 C3 C4 6 1 6 2 0.75 0.142857 0.857143 0.25 Table 10 Examples of numerical values of proportional relationships Amount of data Quadrant 1 Second Quadrant The third quadrant Quadrant 4 C1 C2 C3 C4 6 1 6 2 0.75 0.142857 0.857143 0.25

該處理單元20輸出正比關係至該通訊單元10，該通訊單元10將正比關係輸出至該通訊介面10a，藉以傳輸到外部裝置100。The processing unit 20 outputs the proportional relationship to the communication unit 10 , and the communication unit 10 outputs the proportional relationship to the communication interface 10 a to be transmitted to the external device 100 .

〔反比關係〕〔Inverse relationship〕

反比關係則是多數資料落於第二象限Q2與第四象限Q4，第一象限Q1與第三象限Q3的資料相對較少，圖15參照。The inverse relationship is that most of the data falls in the second quadrant Q2 and the fourth quadrant Q4, and the data in the first quadrant Q1 and the third quadrant Q3 are relatively small, see Figure 15.

該處理單元20係於符合下列條件時判斷該資料關係為反比關係：The processing unit 20 determines that the data relationship is an inverse relationship when the following conditions are met:

C2≥P1 且 C4≥P2C2≥P1 and C4≥P2

其中，C2為第二象限Q2中的資料數量與第二象限Q2及其相鄰之第一象限Q1與第三象限Q3之其中一者的資料數量之總合的比值，P1為一第一預定比值，0.5≤P1≤0.7；Wherein, C2 is the ratio of the amount of data in the second quadrant Q2 to the total amount of data in the second quadrant Q2 and one of the adjacent first quadrant Q1 and third quadrant Q3, P1 is a first predetermined ratio, 0.5≤P1≤0.7;

C4為第四象限Q4中的資料數量與第四象限Q4及其相鄰之第一象限Q1與第三象限Q3之其中另一者的資料數量之總合的比值，P2為一第二預定比值，P2為一第二預定比值，0.5≤P2≤0.7。C4 is the ratio of the amount of data in the fourth quadrant Q4 to the total amount of data in the fourth quadrant Q4 and the other of the adjacent first quadrant Q1 and third quadrant Q3, P2 is a second predetermined ratio, P2 is a second predetermined ratio, 0.5≤P2≤0.7.

本實施例中P1以0.5（50%）為例，P2以0.5（50%）為例，符合上述條件代表第一象限Q1及第四象限Q4中有50%以上的資料是在第四象限Q4，第二象限Q2及第三象限Q3中有50%以上的資料是在第二象限Q2；或者第一象限Q1及第二象限Q2中有50%以上的資料是在第二象限Q2，第三象限Q3及第四象限Q4中有50%以上的資料是在第四象限Q4。In this embodiment, P1 is taken as 0.5 (50%), and P2 is taken as 0.5 (50%). Meeting the above conditions means that more than 50% of the data in the first quadrant Q1 and the fourth quadrant Q4 are in the fourth quadrant Q4, and more than 50% of the data in the second quadrant Q2 and the third quadrant Q3 are in the second quadrant Q2; or more than 50% of the data in the first quadrant Q1 and the second quadrant Q2 are in the second quadrant Q2, and more than 50% of the data in the third quadrant Q3 and the fourth quadrant Q4 are in the fourth quadrant Q4.

假設為表11的情況，在分割原點Os處，C2為第二象限Q2中的資料數量與第二象限Q2及第三象限Q3的資料數量之總合的比值，C4為第四象限Q4中的資料數量與第一象限Q1及第四象限Q4的資料數量之總合的比值。C1為第一象限Q1中的資料數量與第一象限Q1及第四象限Q4的資料數量之總合的比值，C3為第三象限Q3中的資料數量與第二象限Q2及第三象限Q3的資料數量之總合的比值。C2大於P1，C4大於P2，因此，處理單元20判斷為正比關係。更進一步，處理單元20可再判斷C2大於C3且C4大於C1，或C3小於P1且C1小於P2，以進一步確定是反比關係。Assuming the situation of Table 11, at the segmentation origin Os, C2 is the ratio of the amount of data in the second quadrant Q2 to the total amount of data in the second quadrant Q2 and the third quadrant Q3, and C4 is the ratio of the amount of data in the fourth quadrant Q4 to the total amount of data in the first quadrant Q1 and the fourth quadrant Q4. C1 is the ratio of the amount of data in the first quadrant Q1 to the total amount of data in the first quadrant Q1 and the fourth quadrant Q4, and C3 is the ratio of the amount of data in the third quadrant Q3 to the total amount of data in the second quadrant Q2 and the third quadrant Q3. C2 is greater than P1, and C4 is greater than P2, so the processing unit 20 determines that it is a proportional relationship. Furthermore, the processing unit 20 may determine that C2 is greater than C3 and C4 is greater than C1, or C3 is less than P1 and C1 is less than P2, to further determine that it is an inverse proportional relationship.

表11 反比關係的數值示例 資料數量 第一象限 第二象限 第三象限 第四象限 C1 C2 C3 C4 2 6 1 6 0.25 0.857143 0.142857 0.75 Table 11 Examples of numerical values of inverse relationships Amount of data Quadrant 1 Second Quadrant The third quadrant Quadrant 4 C1 C2 C3 C4 2 6 1 6 0.25 0.857143 0.142857 0.75

該處理單元20輸出反比關係至該通訊單元10，該通訊單元10將反比關係輸出至該通訊介面10a，藉以傳輸到外部裝置。The processing unit 20 outputs the inverse relationship to the communication unit 10, and the communication unit 10 outputs the inverse relationship to the communication interface 10a to be transmitted to an external device.

在上述第一較佳實施例中，係以在該搜尋範圍找出一個較佳的分割原點Os，並基於較佳的分割原點Os判斷該些資料的資料關係。於後再提供一第二較佳實施例，可基於該搜尋範圍中的多個分割原點Os分別對該些資料判斷資料關係。In the first preferred embodiment, a better segmentation origin Os is found in the search range, and the data relationship of the data is determined based on the better segmentation origin Os. A second preferred embodiment is provided later, and the data relationship of the data can be determined based on multiple segmentation origins Os in the search range.

本發明第二較佳實施例的資料分析方法，同樣可藉由上述之資料分析裝置1執行，該方法包含圖16所示步驟，其與第一較佳實施例大致相同。The data analysis method of the second preferred embodiment of the present invention can also be executed by the above-mentioned data analysis device 1. The method includes the steps shown in Figure 16, which is roughly the same as the first preferred embodiment.

步驟S21：資料分析裝置1取得一資料集，該資料集包括複數個資料（圖17參照），各該資料包括一第一值與一第二值。本步驟與第一實施例之步驟S11相同。Step S21: The data analysis device 1 obtains a data set, which includes a plurality of data (see FIG. 17 ), each of which includes a first value and a second value. This step is the same as step S11 of the first embodiment.

步驟S22：資料分析裝置1對該些資料設定一搜尋範圍S（圖17參照），其中，該搜尋範圍S具有一下限第一值與一上限第一值，一下限第二值與一上限第二值。本步驟與第一實施例之步驟S12相同。Step S22: The data analysis device 1 sets a search range S for the data (see FIG. 17 ), wherein the search range S has a lower limit first value and an upper limit first value, a lower limit second value and an upper limit second value. This step is the same as step S12 of the first embodiment.

該搜尋範圍S可具有一第一搜尋間隔值與一第二搜尋間隔值，第一搜尋間隔值與第二搜尋間隔值用以在步驟S23中使用。舉例而言，於該第一軸X上，該下限第一值為0，該上限第一值為1，第一搜尋間隔值為0.1；於該第二軸Y上，該下限第二值為0，該上限第二值為1，第二搜尋間隔值為0.1。The search range S may have a first search interval value and a second search interval value, and the first search interval value and the second search interval value are used in step S23. For example, on the first axis X, the lower limit first value is 0, the upper limit first value is 1, and the first search interval value is 0.1; on the second axis Y, the lower limit second value is 0, the upper limit second value is 1, and the second search interval value is 0.1.

步驟S23：資料分析裝置1對該些資料於該搜尋範圍S中搜尋複數個第一分割線L1與複數個第二分割線L2，各該第一分割線L1與各該第二分割線L2相交於一個分割原點Os（圖18參照）。亦即，於該搜尋範圍S搜尋了複數個分割原點Os。各該分割原點Os具有一原點第一值與一原點第二值，各該原點第一值介於該下限第一值與該上限第一值之間，各該原點第二值介於該下限第二值與該上限第二值之間，各該原點第一值與各該原點第二值分別對應各該第一分割線L1與各該第二分割線L2。Step S23: The data analysis device 1 searches for a plurality of first cutting lines L1 and a plurality of second cutting lines L2 in the search range S for the data, and each of the first cutting lines L1 and each of the second cutting lines L2 intersects at a cutting origin Os (see FIG. 18 ). That is, a plurality of cutting origins Os are searched in the search range S. Each of the cutting origins Os has a first origin value and a second origin value, each of the first origin values is between the first lower limit value and the first upper limit value, each of the second origin values is between the second lower limit value and the second upper limit value, and each of the first origin values and each of the second origin values correspond to each of the first cutting lines L1 and each of the second cutting lines L2, respectively.

在該下限第一值至該上限第一值之間，每隔一個第一搜尋間隔值設定一個第一值，在該下限第二值至該上限第二值之間，每隔一個第二搜尋間隔值設定一個第二值。較佳者，最小的第一值不等於下限第一值，最大的第一值不等於上限第一值；最小的第二值不等於下限第二值，最大的第二值不等於上限第二值。各該第一值與各該第二值形成各該分割原點Os的座標。圖18中以81個分割原點Os為例。Between the lower limit first value and the upper limit first value, a first value is set every first search interval value, and between the lower limit second value and the upper limit second value, a second value is set every second search interval value. Preferably, the smallest first value is not equal to the lower limit first value, and the largest first value is not equal to the upper limit first value; the smallest second value is not equal to the lower limit second value, and the largest second value is not equal to the upper limit second value. Each of the first values and each of the second values forms the coordinates of each of the segmentation origins Os. FIG. 18 takes 81 segmentation origins Os as an example.

步驟S24：資料分析裝置1依據各該分割原點Os的第一分割線與第二分割線將該些資料區分為四個象限（第一象限Q1至第四象限Q4），該些象限中分別具有該些資料的一部分。Step S24: the data analysis device 1 divides the data into four quadrants (first quadrant Q1 to fourth quadrant Q4) according to the first dividing line and the second dividing line of each dividing origin Os, and each of the quadrants contains a portion of the data.

步驟S25：資料分析裝置1依據基於各該分割原點Os的各該象限中的資料數量與相鄰的象限中的資料數量的比值判斷該些資料的一資料關係，並輸出基於各該分割原點Os的該資料關係。Step S25: The data analysis device 1 determines a data relationship of the data based on the ratio of the data quantity in each quadrant based on each segmentation origin Os to the data quantity in the adjacent quadrant, and outputs the data relationship based on each segmentation origin Os.

換言之，資料分析裝置1計算每一個分割原點Os的各該象限中的資料數量與相鄰的象限中的資料數量的比值，進而判斷基於每一個分割原點Os的資料關係。較佳者，該處理單元20只判斷四個象限皆有資料的分割原點Os的資料關係。In other words, the data analysis device 1 calculates the ratio of the amount of data in each quadrant of each segmentation origin Os to the amount of data in the adjacent quadrants, and then determines the data relationship based on each segmentation origin Os. Preferably, the processing unit 20 only determines the data relationship of the segmentation origin Os with data in all four quadrants.

判斷各分割原點Os的該資料關係的步驟與步驟S15中判斷水平關係、垂直關係、正比關係與反比關係的步驟相同，The step of determining the data relationship of each segmentation origin Os is the same as the step of determining the horizontal relationship, vertical relationship, proportional relationship and inverse relationship in step S15.

亦即：that is:

該處理單元20係於符合下列條件時判斷該資料關係為第一種水平關係：The processing unit 20 determines that the data relationship is a first level relationship when the following conditions are met:

C1≥P1 且 C2≥P2。C1≥P1 and C2≥P2.

該處理單元20係於符合下列條件時判斷該資料關係為第二種水平關係：The processing unit 20 determines that the data relationship is a second level relationship when the following conditions are met:

C4≥P1 且 C3≥P2。C4≥P1 and C3≥P2.

該處理單元20係於符合下列條件時判斷該資料關係為第一種垂直關係：The processing unit 20 determines that the data relationship is a first vertical relationship when the following conditions are met:

C1’≥P1 且 C4’≥P2。C1’≥P1 and C4’≥P2.

該處理單元20係於符合下列條件時判斷該資料關係為第二種垂直關係：The processing unit 20 determines that the data relationship is a second vertical relationship when the following conditions are met:

C2’≥P1 且 C3’≥P2。C2’≥P1 and C3’≥P2.

該處理單元20係於符合下列條件時判斷該資料關係為正比關係：The processing unit 20 determines that the data relationship is a proportional relationship when the following conditions are met:

C1≥P1 且 C3≥P2。C1≥P1 and C3≥P2.

該處理單元20係於符合下列條件時判斷該資料關係為反比關係：The processing unit 20 determines that the data relationship is an inverse relationship when the following conditions are met:

C2≥P1 且 C4≥P2。C2≥P1 and C4≥P2.

該處理單元20輸出基於各該分割原點Os的資料關係至該通訊單元10，該通訊單元10將基於各該分割原點Os的資料關係輸出至該通訊介面10a，藉以傳輸到外部裝置。較佳者，該處理單元20輸出四個象限皆有資料的分割原點Os的資料關係。The processing unit 20 outputs the data relationship based on each of the segmentation origins Os to the communication unit 10, and the communication unit 10 outputs the data relationship based on each of the segmentation origins Os to the communication interface 10a for transmission to an external device. Preferably, the processing unit 20 outputs the data relationship of the segmentation origins Os having data in all four quadrants.

據上所述，本發明的資料分析裝置及資料分析方法藉由將該些資料集的資料以象限分割的方式判斷資料的資料關係，並輸出到外部裝置，以便後續的利用，例如用於作為機器學習的分類標籤，輔助人工判斷資料關係，藉此，提供一種不同於習用的資料分析方式。不使用複雜的統計方法，可應用於小型的運算裝置，例如用在物聯網的邊緣運算裝置。As described above, the data analysis device and data analysis method of the present invention determine the data relationship of the data by dividing the data sets into quadrants, and output them to an external device for subsequent use, such as being used as classification labels for machine learning to assist manual judgment of data relationships, thereby providing a data analysis method different from the usual one. It does not use complex statistical methods and can be applied to small computing devices, such as edge computing devices used in the Internet of Things.

以上所述僅為本發明較佳可行實施例而已，舉凡應用本發明說明書及申請專利範圍所為之等效變化，理應包含在本發明之專利範圍內。The above description is only the preferred embodiment of the present invention. Any equivalent changes made by applying the present invention specification and the scope of patent application should be included in the patent scope of the present invention.

1:資料分析裝置1: Data analysis device

10:通訊單元10: Communication unit

10a:通訊介面10a: Communication Interface

20:處理單元20: Processing unit

30:儲存單元30: Storage unit

100:外部裝置100: External device

Os:分割原點Os: Split origin

L1:第一分割線L1: First dividing line

L2:第二分割線L2: The second dividing line

Q1:第一象限Q1: Quadrant 1

Q2:第二象限Q2: Second Quadrant

Q3:第三象限Q3: The Third Quadrant

Q4:第四象限Q4: The fourth quadrant

O、O1、O3、O4:搜尋原點O, O1, O3, O4: Search for origin

O2、O5、O6、O7:搜尋原點(參考原點)O2, O5, O6, O7: Search origin (reference origin)

Lp1:第一預定分割線Lp1: The first predetermined dividing line

Lp2:第二預定分割線Lp2: The second predetermined dividing line

Qp1:第一預定象限Qp1: First predetermined quadrant

Qp2:第二預定象限Qp2: Second predetermined quadrant

Qp3:第三預定象限Qp3: The third predetermined quadrant

Qp4:第四預定象限Qp4: The fourth predetermined quadrant

Lr1:第一參考分割線Lr1: First reference dividing line

Lr2:第二參考分割線Lr2: Second reference dividing line

Qr1:第一參考象限Qr1: First reference quadrant

Qr2:第二參考象限Qr2: Second reference quadrant

Qr3:第三參考象限Qr3: The third reference quadrant

Qr4:第四參考象限Qr4: Fourth reference quadrant

M:分界線M: Dividing Line

Lh:水平分割線Lh: horizontal dividing line

Lv:垂直分割線Lv: vertical dividing line

S:搜尋範圍S: Search range

S11~S15:步驟S11~S15: Steps

X:第一軸X: First axis

Y:第二軸Y: Second axis

圖1本發明一第一較佳實施例之資料分析裝置的示意圖。 圖2為本發明上述較佳實施例之資料分析方法的流程圖。 圖3為本發明上述較佳實施例之資料的第一種水平關係的分布示意圖。 圖4為本發明上述較佳實施例設定搜尋原點的示意圖。 圖5為本發明上述較佳實施例以搜尋原點分成第一至第四預定象限之示意圖。 圖6為本發明上述較佳實施例以搜尋原點形成分割原點之示意圖。 圖7為本發明以多個搜尋原點連成分界線之示意圖。 圖8為本發明上述較佳實施例產生水平分割線之示意圖。 圖9為本發明上述較佳實施例資料的第二種水平關係的分布及產生水平分割線之示意圖。 圖10為本發明上述較佳實施例資料的第一種垂直關係的分布示意圖。 圖11為本發明上述較佳實施例資料的第一種垂直關係的分布及產生垂直分割線之示意圖。 圖12為本發明上述較佳實施例資料的第二種垂直關係的分布示意圖。 圖13為本發明上述較佳實施例資料的第二種垂直關係的分布及產生垂直分割線之示意圖。 圖14為本發明上述較佳實施例資料的正比關係的分布示意圖。 圖15為本發明上述較佳實施例資料的反比關係的分布示意圖。 圖16為本發明第二較佳實施例之資料分析方法的流程圖。 圖17為本發明上述較佳實施例之資料的分布示意圖。 圖18為本發明上述較佳實施例設定分割原點的示意圖。 FIG1 is a schematic diagram of a data analysis device of a first preferred embodiment of the present invention. FIG2 is a flow chart of a data analysis method of the preferred embodiment of the present invention. FIG3 is a schematic diagram of the distribution of the first horizontal relationship of the data of the preferred embodiment of the present invention. FIG4 is a schematic diagram of setting the search origin of the preferred embodiment of the present invention. FIG5 is a schematic diagram of the preferred embodiment of the present invention divided into the first to fourth predetermined quadrants by the search origin. FIG6 is a schematic diagram of the preferred embodiment of the present invention forming a segmentation origin by the search origin. FIG7 is a schematic diagram of the present invention connecting a plurality of search origins to form a dividing line. FIG8 is a schematic diagram of the preferred embodiment of the present invention generating a horizontal dividing line. Figure 9 is a schematic diagram of the distribution of the second horizontal relationship of the above-mentioned preferred embodiment data of the present invention and the generation of horizontal dividing lines. Figure 10 is a schematic diagram of the distribution of the first vertical relationship of the above-mentioned preferred embodiment data of the present invention. Figure 11 is a schematic diagram of the distribution of the first vertical relationship of the above-mentioned preferred embodiment data of the present invention and the generation of vertical dividing lines. Figure 12 is a schematic diagram of the distribution of the second vertical relationship of the above-mentioned preferred embodiment data of the present invention. Figure 13 is a schematic diagram of the distribution of the second vertical relationship of the above-mentioned preferred embodiment data of the present invention and the generation of vertical dividing lines. Figure 14 is a schematic diagram of the distribution of the proportional relationship of the above-mentioned preferred embodiment data of the present invention. Figure 15 is a schematic diagram of the distribution of the inverse relationship of the above-mentioned preferred embodiment data of the present invention. FIG. 16 is a flow chart of the data analysis method of the second preferred embodiment of the present invention. FIG. 17 is a schematic diagram of the distribution of data of the above preferred embodiment of the present invention. FIG. 18 is a schematic diagram of setting the segmentation origin of the above preferred embodiment of the present invention.

S11~S15:步驟 S11~S15: Steps

Claims (18)

一種資料分析方法，係應用於一資料分析裝置，該資料分析方法包含下列步驟：A.該資料分析裝置取得一資料集，該資料集包括複數個資料，各該資料包括一第一值與一第二值；B.該資料分析裝置對該些資料設定一搜尋範圍，其中，該搜尋範圍具有一下限第一值與一上限第一值，一下限第二值與一上限第二值；C.該資料分析裝置對該些資料於該搜尋範圍中搜尋至少一第一分割線與至少一第二分割線，該至少一第一分割線與該至少一第二分割線相交於至少一分割原點，該至少一分割原點具有一原點第一值與一原點第二值，該原點第一值介於該下限第一值與該上限第一值之間，該原點第二值介於該下限第二值與該上限第二值之間，該原點第一值與該原點第二值分別對應該至少一第一分割線與該至少一第二分割線；D.該資料分析裝置依據該至少一分割原點的該至少一第一分割線與該至少一第二分割線將該些資料區分為四個象限，該些象限中分別具有該些資料的一部分；E.該資料分析裝置依據基於該至少一分割原點的各該象限中的資料數量與相鄰的象限中的資料數量的比值判斷該些資料的一資料關係，並輸出基於該至少一分割原點的該資料關係。 A data analysis method is applied to a data analysis device, and the data analysis method includes the following steps: A. The data analysis device obtains a data set, and the data set includes a plurality of data, each of which includes a first value and a second value; B. The data analysis device sets a search range for the data, wherein the search range has a lower limit first value and an upper limit first value, and a lower limit second value and an upper limit second value; C. The data analysis device searches for at least one first dividing line and at least one second dividing line in the search range for the data, and the at least one first dividing line and the at least one second dividing line intersect at at least one dividing origin, and the at least one dividing origin has a first origin value and a second origin value, and the The first value of the origin is between the first value of the lower limit and the first value of the upper limit, the second value of the origin is between the second value of the lower limit and the second value of the upper limit, and the first value of the origin and the second value of the origin correspond to the at least one first dividing line and the at least one second dividing line respectively; D. The data analysis device divides the data into four quadrants according to the at least one first dividing line and the at least one second dividing line of the at least one dividing origin, and the quadrants respectively have a part of the data; E. The data analysis device determines a data relationship of the data according to the ratio of the amount of data in each quadrant based on the at least one dividing origin to the amount of data in the adjacent quadrant, and outputs the data relationship based on the at least one dividing origin. 如請求項1所述之資料分析方法，其中步驟C中該至少一第一分割線、該至少一第二分割線與該至少一分割原點的數量為複數個；步驟D中，係依據各該分割原點的第一分割線與第二分割線將該些資料區分為四個象限； 步驟E中，係依據基於各該分割原點的資料數量與相鄰的象限中的資料數量的比值判斷該些資料的該資料關係，並輸出基於各該分割原點的該資料關係。 The data analysis method as described in claim 1, wherein in step C, the number of the at least one first dividing line, the at least one second dividing line and the at least one dividing origin is plural; in step D, the data are divided into four quadrants according to the first dividing line and the second dividing line of each dividing origin; in step E, the data relationship of the data is determined according to the ratio of the data quantity based on each dividing origin to the data quantity in the adjacent quadrant, and the data relationship based on each dividing origin is output. 如請求項1所述之資料分析方法，其中步驟C中搜尋該至少一第一分割線與該至少一第二分割線的步驟包含：該資料分析裝置設定複數個搜尋原點，各該搜尋原點具有一預定第一值與一預定第二值，各該預定第一值介於該下限第一值與該上限第一值之間，各該預定第二值介於該下限第二值與該上限第二值之間；該資料分析裝置以各該搜尋原點的預定第一值與預定第二值分別形成的一第一預定分割線與一第二預定分割線將該些資料區分為四個預定象限，該四個預定象限分別為一第一預定象限、一第二預定象限、一第三預定象限與一第四預定象限；該資料分析裝置分別計算基於各該搜尋原點的Z1y、Z2y、Z3y、Z4y、Z5y、Z6y、Z1x、Z2x、Z3x、Z4x、Z5x與Z6x數值，其中：其中，Z1y為C1y與C2y之中的最小值；Z2y為C3y與C4y之中的最小值；Z3y為C1y與C4y之中的最小值；Z4y為C2y與C3y之中的最小值；Z5y為C1y與C3y之中的最小值；Z6y為C2y與C4y之中的最小值；Z1x為C1x與C2x之中的最小值；Z2x為C3x與C4x之中的最小值；Z3x為C1x與C4x之中的最小值；Z4x為C2x與C3x之中的最小值；Z5x為C1x與C3x之中的最小值；Z6x為C2x與C4x之中的最小值；其中，C1y為該第一預定象限中的資料數量與該第一預定象限及該第二預定象限中的資料數量之總合的比值；C2y為該第二預定象限中的資料數量與該第一預定象限及該第二預定象限中的資料數量之總合的比值；C3y為該第三預定象限中的資料數量與該第三預定象限及該第四預 定象限中的資料數量之總合的比值；C4y為該第四預定象限中的資料數量與該第三預定象限及該第四預定象限中的資料數量之總合的比值；C1x為該第一預定象限中的資料數量與該第一預定象限及該第四預定象限中的資料數量之總合的比值；C2x為該第二預定象限中的資料數量與該第二預定象限及該第三預定象限中的資料數量之總合的比值；C3x為該第三預定象限中的資料數量與該第二預定象限及該第三預定象限中的資料數量之總合的比值；C4x為該第四預定象限中的資料數量與該第一預定象限及該第四預定象限中的資料數量之總合的比值；該資料分析裝置由該些搜尋原點之Z1y、Z1x、Z2y、Z2x、Z3y、Z3x、Z4y、Z4x、Z5y、Z5x、Z6y與Z6x數值中取最大者的搜尋原點據以設定該至少一分割原點。 The data analysis method as described in claim 1, wherein the step of searching for the at least one first dividing line and the at least one second dividing line in step C comprises: the data analysis device sets a plurality of search origins, each of the search origins has a predetermined first value and a predetermined second value, each of the predetermined first values is between the lower limit first value and the upper limit first value, and each of the predetermined second values is between the lower limit second value and the upper limit second value; the data analysis device divides the data into four predetermined quadrants using a first predetermined dividing line and a second predetermined dividing line respectively formed by the predetermined first value and the predetermined second value of each of the search origins, the four predetermined quadrants being a first predetermined quadrant, a second predetermined quadrant, a third predetermined quadrant and a fourth predetermined quadrant; the data analysis device The values of Z1y, Z2y, Z3y, Z4y, Z5y, Z6y, Z1x, Z2x, Z3x, Z4x, Z5x and Z6x based on the search origin are calculated respectively, wherein: Z1y is the minimum value between C1y and C2y; Z2y is the minimum value between C3y and C4y; Z3y is the minimum value between C1y and C4y; Z4y is the minimum value between C2y and C3y; Z5y is the minimum value between C1y and C3y; Z6y is the minimum value between C2y and C4y; Z1x is the minimum value between C1x and C2x; Z2x is the minimum value between C3x and C4x; Z3x is the minimum value between C1x and C4x; Z4x is the minimum value between C2x and C3x ; Z5x is the minimum value between C1x and C3x; Z6x is the minimum value between C2x and C4x; wherein, C1y is the ratio of the amount of data in the first predetermined quadrant to the sum of the amount of data in the first predetermined quadrant and the second predetermined quadrant; C2y is the ratio of the amount of data in the second predetermined quadrant to the sum of the amount of data in the first predetermined quadrant and the second predetermined quadrant; C3y is the ratio of the amount of data in the third predetermined quadrant to the sum of the amount of data in the third predetermined quadrant and the fourth predetermined quadrant; C4y is the ratio of the amount of data in the fourth predetermined quadrant to the sum of the amount of data in the third predetermined quadrant and the fourth predetermined quadrant; C1x is the ratio of the amount of data in the first predetermined quadrant to the sum of the amount of data in the first predetermined quadrant. quadrant and the fourth predetermined quadrant; C2x is the ratio of the amount of data in the second predetermined quadrant to the sum of the amount of data in the second predetermined quadrant and the third predetermined quadrant; C3x is the ratio of the amount of data in the third predetermined quadrant to the sum of the amount of data in the second predetermined quadrant and the third predetermined quadrant; C4x is the ratio of the amount of data in the fourth predetermined quadrant to the sum of the amount of data in the first predetermined quadrant and the fourth predetermined quadrant; the data analysis device sets the at least one segmentation origin based on the search origin that takes the largest value among the values of Z1y, Z1x, Z2y, Z2x, Z3y, Z3x, Z4y, Z4x, Z5y, Z5x, Z6y and Z6x of the search origins. 如請求項3所述之資料分析方法，其中，該些搜尋原點之Z1y、Z1x、Z2y、Z2x、Z3y、Z3x、Z4y、Z4x、Z5y、Z5x、Z6y與Z6x中取最大者的搜尋原點為複數個時，則將複數個該搜尋原點的預定第一值平均及預定第二值平均後分別設定為該分割原點的原點第一值與一原點第二值。 The data analysis method as described in claim 3, wherein when there are multiple search origins that take the largest one among the search origins Z1y, Z1x, Z2y, Z2x, Z3y, Z3x, Z4y, Z4x, Z5y, Z5x, Z6y and Z6x, the predetermined first values and the predetermined second values of the multiple search origins are averaged and set as the origin first value and the origin second value of the split origin respectively. 如請求項4所述之資料分析方法，其中，該些搜尋原點之Z1y、Z1x、Z2y、Z2x、Z3y、Z3x、Z4y、Z4x、Z5y、Z5x、Z6y與Z6x中取最大者的搜尋原點為複數個時，則基於對應的複數個該搜尋原點形成一分界線。 The data analysis method as described in claim 4, wherein when there are multiple search origins that are the largest among the search origins Z1y, Z1x, Z2y, Z2x, Z3y, Z3x, Z4y, Z4x, Z5y, Z5x, Z6y and Z6x, a dividing line is formed based on the corresponding multiple search origins. 如請求項3所述之資料分析方法，其中步驟D中該四個象限分別為一第一象限、一第二象限、一第三象限與一第四象限；其中步驟E中係於符合下列條件時判斷該資料關係為一水平關係：C1

P1且C2

P2； 其中，C1為該第一象限中的資料數量與該第一象限及該第四象限中的資料數量之總合的比值，P1為一第一預定比值；C2為該第二象限中的資料數量與該第二象限及該第三象限中的資料數量之總合的比值，P2為一第二預定比值。 The data analysis method as described in claim 3, wherein in step D, the four quadrants are a first quadrant, a second quadrant, a third quadrant, and a fourth quadrant; and in step E, the data relationship is determined to be a horizontal relationship when the following conditions are met: C1

P1 and C2

P2; wherein C1 is the ratio of the amount of data in the first quadrant to the total amount of data in the first quadrant and the fourth quadrant, and P1 is a first predetermined ratio; C2 is the ratio of the amount of data in the second quadrant to the total amount of data in the second quadrant and the third quadrant, and P2 is a second predetermined ratio. 如請求項6所述之資料分析方法，更包含下列步驟：該資料分析裝置選定該原點第一值，並且由該下限第二值與該上限第二值之間選定與該原點第一值配合的複數個參考第二值，以形成複數個參考原點，各該參考原點具有一參考第一值，各該參考第一值為該原點第一值；該資料分析裝置以各該參考原點的參考第一值與參考第二值分別形成一第一參考分割線與一第二參考分割線，且各該參考原點的第一參考分割線與第二參考分割線將該些資料區分為四個參考象限，該四個參考象限分別為一第一參考象限、一第二參考象限、一第三參考象限與一第四參考象限；該資料分析裝置分別計算基於各該參考原點的Csum1數值，其中：Csum1=C11x+C21x；其中，C11x為該第一參考象限中的資料數量與該第一參考象限及該第四參考象限中的資料數量之總合的比值；C21x為該第二參考象限中的資料數量與該第二參考象限及該第三參考象限中的資料數量之總合的比值；該資料分析裝置計算各該參考原點的Csum1與其相鄰且參考第二值較大之參考原點的Csum1之差值，依據具有最大之差值的參考原點的參考第二值設定一水平分割線。 The data analysis method as described in claim 6 further includes the following steps: the data analysis device selects the first value of the origin, and selects a plurality of reference second values between the lower limit second value and the upper limit second value that match the first value of the origin to form a plurality of reference origins, each of the reference origins having a reference first value, and each of the reference first values is the first value of the origin; the data analysis device forms a first reference dividing line and a second reference dividing line with the reference first value and the reference second value of each of the reference origins, respectively, and the first reference dividing line and the second reference dividing line of each of the reference origins divide the data into four reference quadrants, and the four reference quadrants are a first reference quadrant, a second reference quadrant, and a third quadrant, respectively. reference quadrant and a fourth reference quadrant; the data analysis device calculates the Csum1 value based on each reference origin respectively, wherein: Csum1=C11x+C21x; wherein, C11x is the ratio of the amount of data in the first reference quadrant to the sum of the amount of data in the first reference quadrant and the fourth reference quadrant; C21x is the ratio of the amount of data in the second reference quadrant to the sum of the amount of data in the second reference quadrant and the third reference quadrant; the data analysis device calculates the difference between the Csum1 of each reference origin and the Csum1 of the adjacent reference origin with a larger reference second value, and sets a horizontal dividing line according to the reference second value of the reference origin with the largest difference. 如請求項3所述之資料分析方法，其中步驟D中該四個象限分別為一第一象限、一第二象限、一第三象限與一第四象限；其中步驟E中係於符合下列條件時判斷該資料關係為一水平關係：C4

P1且C3

P2；其中，C4為該第四象限中的資料數量與該第一象限及該第四象限中的資料數量之總合的比值，P1為一第一預定比值；C3為該第三象限中的資料數量與該第二象限及該第三象限中的資料數量之總合的比值，P2為一第二預定比值。 The data analysis method as described in claim 3, wherein in step D, the four quadrants are a first quadrant, a second quadrant, a third quadrant, and a fourth quadrant; wherein in step E, the data relationship is determined to be a horizontal relationship when the following conditions are met: C4

P1 and C3

P2; wherein, C4 is the ratio of the amount of data in the fourth quadrant to the total amount of data in the first quadrant and the fourth quadrant, and P1 is a first predetermined ratio; C3 is the ratio of the amount of data in the third quadrant to the total amount of data in the second quadrant and the third quadrant, and P2 is a second predetermined ratio. 如請求項8所述之資料分析方法，更包含下列步驟：該資料分析裝置選定該原點第一值，並且由該下限第二值與該上限第二值之間選定與該原點第一值配合的複數個參考第二值，以形成複數個參考原點，各該參考原點具有一參考第一值，各該參考第一值為該原點第一值；該資料分析裝置以各該參考原點的參考第一值與參考第二值分別形成一第一參考分割線與一第二參考分割線，且各該參考原點的第一參考分割線與第二參考分割線將該些資料區分為四個參考象限，該四個參考象限分別為一第一參考象限、一第二參考象限、一第三參考象限與一第四參考象限；該資料分析裝置分別計算基於各該參考原點的Csum2數值，其中：Csum2=C31x+C41x；其中，C31x為該第三參考象限中的資料數量與該第二參考象限及該第三參考象限中的資料數量之總合的比值；C41x為該第四參考象限中的資料數量與該第一參考象限及該第四參考象限中的資料數量之總合的比值； 該資料分析裝置計算各該參考原點的Csum2與其相鄰且參考第二值較小之參考原點的Csum2之差值，依據具有最大之差值的參考原點的參考第二值設定一水平分割線。 The data analysis method as described in claim 8 further includes the following steps: the data analysis device selects the first value of the origin, and selects a plurality of reference second values between the lower limit second value and the upper limit second value that match the first value of the origin to form a plurality of reference origins, each of the reference origins having a reference first value, and each of the reference first values is the first value of the origin; the data analysis device forms a first reference dividing line and a second reference dividing line with the reference first value and the reference second value of each of the reference origins, respectively, and the first reference dividing line and the second reference dividing line of each of the reference origins divide the data into four reference quadrants, and the four reference quadrants are a first reference quadrant, a second reference quadrant, and a third reference quadrant, respectively. quadrant and a fourth reference quadrant; the data analysis device calculates the Csum2 value based on each reference origin respectively, where: Csum2=C31x+C41x; where C31x is the ratio of the amount of data in the third reference quadrant to the sum of the amount of data in the second reference quadrant and the third reference quadrant; C41x is the ratio of the amount of data in the fourth reference quadrant to the sum of the amount of data in the first reference quadrant and the fourth reference quadrant; The data analysis device calculates the difference between the Csum2 of each reference origin and the Csum2 of the adjacent reference origin with a smaller reference second value, and sets a horizontal dividing line according to the reference second value of the reference origin with the largest difference. 如請求項3所述之資料分析方法，其中步驟D中該四個象限分別為一第一象限、一第二象限、一第三象限與一第四象限；其中步驟E中係於符合下列條件時判斷該資料關係為一垂直關係：C1’

P1且C4’

P2；其中，C1’為該第一象限中的資料數量與該第一象限及該第二象限中的資料數量之總合的比值，P1為一第一預定比值；C4’為該第四象限中的資料數量與該第三象限及該第四象限中的資料數量之總合的比值，P2為一第二預定比值。 The data analysis method as described in claim 3, wherein the four quadrants in step D are a first quadrant, a second quadrant, a third quadrant, and a fourth quadrant; wherein in step E, the data relationship is determined to be a vertical relationship when the following conditions are met: C1'

P1 and C4'

P2; wherein, C1' is the ratio of the amount of data in the first quadrant to the total amount of data in the first quadrant and the second quadrant, and P1 is a first predetermined ratio; C4' is the ratio of the amount of data in the fourth quadrant to the total amount of data in the third quadrant and the fourth quadrant, and P2 is a second predetermined ratio. 如請求項10所述之資料分析方法，更包含下列步驟：該資料分析裝置選定該原點第二值，並且由該下限第一值與該上限第一值之間選定與該原點第二值配合的複數個參考第一值，以形成複數個參考原點，各該參考原點具有一參考第二值，各該參考第二值為該原點第二值；該資料分析裝置以各該參考原點的參考第一值與參考第二值分別形成一第一參考分割線與一第二參考分割線，且各該參考原點的第一參考分割線與第二參考分割線將該些資料區分為四個參考象限，該四個參考象限分別為一第一參考象限、一第二參考象限、一第三參考象限與一第四參考象限；該資料分析裝置分別計算基於各該參考原點的Csum1’數值，其中：Csum1’=C11y+C41y； 其中，C11y為該第一參考象限中的資料數量與該第一參考象限及該第二參考象限中的資料數量之總合的比值；C41y為該第四參考象限中的資料數量與該第三參考象限及該第四參考象限中的資料數量之總合的比值；該資料分析裝置計算各該參考原點的Csum1’與其相鄰且參考第一值較大之參考原點的Csum1’之差值，依據具有最大之差值的參考原點的參考第一值設定一垂直分割線。 The data analysis method as described in claim 10 further comprises the following steps: the data analysis device selects the second value of the origin, and selects a plurality of reference first values matching the second value of the origin between the lower limit first value and the upper limit first value to form a plurality of reference origins, each of which has a reference second value, and each of which is the second value of the origin; the data analysis device forms a first reference dividing line and a second reference dividing line with the reference first value and the reference second value of each of which, respectively, and the first reference dividing line and the second reference dividing line of each of which divide the data into four reference quadrants, and the four reference quadrants are a first reference quadrant, a second reference quadrant, a third reference quadrant, and a fourth quadrant. quadrant and a fourth reference quadrant; the data analysis device calculates the Csum1' value based on each reference origin respectively, wherein: Csum1'=C11y+C41y; wherein, C11y is the ratio of the amount of data in the first reference quadrant to the sum of the amount of data in the first reference quadrant and the second reference quadrant; C41y is the ratio of the amount of data in the fourth reference quadrant to the sum of the amount of data in the third reference quadrant and the fourth reference quadrant; the data analysis device calculates the difference between the Csum1' of each reference origin and the Csum1' of the adjacent reference origin with a larger reference first value, and sets a vertical dividing line according to the reference first value of the reference origin with the largest difference. 如請求項3所述之資料分析方法，其中步驟D中該四個象限分別為一第一象限、一第二象限、一第三象限與一第四象限；其中步驟E中係於符合下列條件時判斷該資料關係為一垂直關係：C2’

P1且C3’

P2；其中，C2’為該第二象限中的資料數量與該第一象限及該第二象限中的資料數量之總合的比值，P1為一第一預定比值；C3’為該第三象限中的資料數量與該第三象限及該第四象限中的資料數量之總合的比值，P2為一第二預定比值。 The data analysis method as described in claim 3, wherein in step D, the four quadrants are a first quadrant, a second quadrant, a third quadrant, and a fourth quadrant; wherein in step E, the data relationship is determined to be a vertical relationship when the following conditions are met: C2'

P1 and C3'

P2; wherein, C2' is the ratio of the amount of data in the second quadrant to the total amount of data in the first quadrant and the second quadrant, and P1 is a first predetermined ratio; C3' is the ratio of the amount of data in the third quadrant to the total amount of data in the third quadrant and the fourth quadrant, and P2 is a second predetermined ratio. 如請求項12所述之資料分析方法，更包含下列步驟：該資料分析裝置選定該原點第二值，並且由該下限第一值與該上限第一值之間選定與該原點第二值配合的複數個參考第一值，以形成複數個參考原點，各該參考原點具有一參考第二值，各該參考第二值為該原點第二值；該資料分析裝置以各該參考原點的參考第一值與參考第二值分別形成一第一參考分割線與一第二參考分割線，且各該參考原點的第一參考分割線與第二參考分割線將該些資料區分為四個參考象限，該四個參考 象限分別為一第一參考象限、一第二參考象限、一第三參考象限與一第四參考象限；該資料分析裝置分別計算基於各該參考原點的Csum2’數值，其中：Csum2’=C21y+C31y；其中，C21y為該第二參考象限中的資料數量與該第一參考象限及該第二參考象限中的資料數量之總合的比值；C31y為該第三參考象限中的資料數量與該第三參考象限及該第四參考象限中的資料數量之總合的比值；該資料分析裝置計算各該參考原點的Csum2’與其相鄰且參考第二值較小之參考原點的Csum2’之差值，依據具有最大之差值的參考原點的參考第二值設定一垂直分割線。 The data analysis method as described in claim 12 further comprises the following steps: the data analysis device selects the second value of the origin, and selects a plurality of reference first values that match the second value of the origin from between the lower limit first value and the upper limit first value to form a plurality of reference origins, each of which has a reference second value, and each of which is the second value of the origin; the data analysis device forms a first reference dividing line and a second reference dividing line with the reference first value and the reference second value of each reference origin, and the first reference dividing line and the second reference dividing line of each reference origin divide the data into four reference quadrants, and the four reference quadrants are a first reference quadrant, a second reference quadrant, a third reference quadrant, and a fourth reference quadrant. quadrant and a fourth reference quadrant; the data analysis device calculates the Csum2' value based on each reference origin respectively, where: Csum2'=C21y+C31y; where C21y is the ratio of the amount of data in the second reference quadrant to the sum of the amount of data in the first reference quadrant and the second reference quadrant; C31y is the ratio of the amount of data in the third reference quadrant to the sum of the amount of data in the third reference quadrant and the fourth reference quadrant; the data analysis device calculates the difference between the Csum2' of each reference origin and the Csum2' of the adjacent reference origin with a smaller reference second value, and sets a vertical dividing line according to the reference second value of the reference origin with the largest difference. 如請求項1所述之資料分析方法，其中步驟D中該四個象限分別為一第一象限、一第二象限、一第三象限與一第四象限；其中，步驟E中係於符合下列條件時判斷該資料關係為一第一水平關係：C1

P1且C2

P2；其中，C1為該第一象限中的資料數量與該第一象限及該第四象限中的資料數量之總合的比值，P1為一第一預定比值；C2為該第二象限中的資料數量與該第二象限及該第三象限中的資料數量之總合的比值，P2為一第二預定比值；其中，步驟E中係於符合下列條件時判斷該資料關係為一第二水平關係：C4

P1且C3

P2； 其中，C4為該第四象限中的資料數量與該第一象限及該第四象限中的資料數量之總合的比值；C3為該第三象限中的資料數量與該第二象限及該第三象限中的資料數量之總合的比值。 The data analysis method as described in claim 1, wherein the four quadrants in step D are a first quadrant, a second quadrant, a third quadrant, and a fourth quadrant; wherein in step E, the data relationship is determined to be a first level relationship when the following conditions are met: C1

P1 and C2

P2; wherein C1 is the ratio of the amount of data in the first quadrant to the total amount of data in the first quadrant and the fourth quadrant, and P1 is a first predetermined ratio; C2 is the ratio of the amount of data in the second quadrant to the total amount of data in the second quadrant and the third quadrant, and P2 is a second predetermined ratio; wherein, in step E, the data relationship is determined to be a second level relationship when the following conditions are met: C4

P1 and C3

P2; wherein C4 is the ratio of the amount of data in the fourth quadrant to the total amount of data in the first quadrant and the fourth quadrant; and C3 is the ratio of the amount of data in the third quadrant to the total amount of data in the second quadrant and the third quadrant. 如請求項1所述之資料分析方法，其中步驟D中該四個象限分別為一第一象限、一第二象限、一第三象限與一第四象限；其中，步驟E中係於符合下列條件時判斷該資料關係為一第一垂直關係：C1’

P1且C4’

P2；其中，C1’為該第一象限中的資料數量與該第一象限及該第二象限中的資料數量之總合的比值，P1為一第一預定比值；C4’為該第四象限中的資料數量與該第三象限及該第四象限中的資料數量之總合的比值，P2為一第二預定比值；其中，步驟E中係於符合下列條件時判斷該資料關係為一第二垂直關係：C2’

P1且C3’

P2；其中，C2’為該第二象限中的資料數量與該第一象限及該第二象限中的資料數量之總合的比值；C3’為該第三象限中的資料數量與該第三象限及該第四象限中的資料數量之總合的比值。 The data analysis method as described in claim 1, wherein the four quadrants in step D are a first quadrant, a second quadrant, a third quadrant, and a fourth quadrant; wherein in step E, the data relationship is determined to be a first vertical relationship when the following conditions are met: C1'

P1 and C4'

P2; wherein C1' is the ratio of the amount of data in the first quadrant to the total amount of data in the first quadrant and the second quadrant, and P1 is a first predetermined ratio; C4' is the ratio of the amount of data in the fourth quadrant to the total amount of data in the third quadrant and the fourth quadrant, and P2 is a second predetermined ratio; wherein, in step E, the data relationship is determined to be a second vertical relationship when the following conditions are met: C2'

P1 and C3'

P2; wherein C2' is the ratio of the amount of data in the second quadrant to the total amount of data in the first quadrant and the second quadrant; and C3' is the ratio of the amount of data in the third quadrant to the total amount of data in the third quadrant and the fourth quadrant. 如請求項1所述之資料分析方法，其中步驟D中該四個象限分別為一第一象限、一第二象限、一第三象限與一第四象限；其中步驟E中係於符合下列條件時判斷該資料關係為一正比關係：C1

P1且C3

P2；其中，C1為該第一象限中的資料數量與該第一象限及其相鄰之該第二象限與該第四象限之其中一者的資料數量之總合的比值，P1為一第一 預定比值；C3為該第三象限中的資料數量與該第三象限及其相鄰之該第二象限與該第四象限之其中另一者的資料數量之總合的比值，P2為一第二預定比值。 The data analysis method as described in claim 1, wherein in step D, the four quadrants are a first quadrant, a second quadrant, a third quadrant, and a fourth quadrant; wherein in step E, the data relationship is determined to be a proportional relationship when the following conditions are met: C1

P1 and C3

P2; wherein, C1 is the ratio of the amount of data in the first quadrant to the total amount of data in the first quadrant and one of the adjacent second quadrant and the fourth quadrant, and P1 is a first predetermined ratio; C3 is the ratio of the amount of data in the third quadrant to the total amount of data in the third quadrant and the other of the adjacent second quadrant and the fourth quadrant, and P2 is a second predetermined ratio. 如請求項1所述之資料分析方法，其中步驟D中該四個象限分別為一第一象限、一第二象限、一第三象限與一第四象限；其中步驟E中係於符合下列條件時判斷該資料關係為一反比關係：C2

P1且C4

P2；其中，C2為該第二象限中的資料數量與該第二象限及其相鄰之該第一象限與該第三象限之其中一者的資料數量之總合的比值，P1為一第一預定比值；C4為該第四象限中的資料數量與該第四象限及其相鄰之該第一象限與該第三象限之其中另一者的資料數量之總合的比值，P2為一第二預定比值。 The data analysis method as described in claim 1, wherein in step D, the four quadrants are a first quadrant, a second quadrant, a third quadrant, and a fourth quadrant; wherein in step E, the data relationship is determined to be an inverse relationship when the following conditions are met: C2

P1 and C4

P2; wherein, C2 is the ratio of the amount of data in the second quadrant to the total amount of data in the second quadrant and one of the adjacent first quadrant and the third quadrant, and P1 is a first predetermined ratio; C4 is the ratio of the amount of data in the fourth quadrant to the total amount of data in the fourth quadrant and the other of the adjacent first quadrant and the third quadrant, and P2 is a second predetermined ratio. 一種資料分析裝置，包含：一通訊單元，供連接一通訊介面且自該通訊介面接收一資料集，該資料集包括複數個資料，各該資料包括一第一值與一第二值；一處理單元，電性連接該通訊單元，該處理單元取得該資料集，並且對該些資料於一搜尋範圍中搜尋至少一第一分割線與至少一第二分割線，其中，該搜尋範圍具有一下限第一值與一上限第一值，一下限第二值與一上限第二值；該至少一第一分割線與該至少一第二分割線相交於至少一分割原點，該至少一分割原點具有一原點第一值與一原點第二值，該原點第一值介於該下限第一值與該上限第一值之間，該原點第二值介於該下限第二值與該上限第二值之間；其中，該處理單元依據該至少一第一分割線與該至少一第二分割線將該些資料區分為四個象限，該些象限中分別具有該些資料的一部分； 該處理單元依據基於該至少一分割原點的各該象限中的資料數量與相鄰的象限中的資料數量的比值判斷該些資料的一資料關係，並輸出基於該至少一分割原點的該資料關係至該通訊單元；其中，該通訊單元將基於該至少一分割原點的該資料關係輸出至該通訊介面。 A data analysis device comprises: a communication unit for connecting to a communication interface and receiving a data set from the communication interface, the data set comprising a plurality of data, each of which comprises a first value and a second value; a processing unit electrically connected to the communication unit, the processing unit obtaining the data set and searching for at least one first dividing line and at least one second dividing line in a search range for the data, wherein the search range has a lower limit first value and an upper limit first value, a lower limit second value and an upper limit second value; the at least one first dividing line and the at least one second dividing line intersect at at least one dividing origin, the at least one dividing origin has a first origin value and a second origin value, the origin The first value of the point is between the first value of the lower limit and the first value of the upper limit, and the second value of the origin is between the second value of the lower limit and the second value of the upper limit; wherein the processing unit divides the data into four quadrants according to the at least one first dividing line and the at least one second dividing line, and each of the quadrants has a part of the data; The processing unit determines a data relationship of the data according to the ratio of the amount of data in each quadrant based on the at least one dividing origin to the amount of data in the adjacent quadrant, and outputs the data relationship based on the at least one dividing origin to the communication unit; wherein the communication unit outputs the data relationship based on the at least one dividing origin to the communication interface.

Images