CN114330578A

CN114330578A - Object classification method and device, readable medium and electronic equipment

Info

Publication number: CN114330578A
Application number: CN202111671141.2A
Authority: CN
Inventors: 钱炜烁; 石崇文; 潘煜文; 苏博览; 黄博
Original assignee: Beijing Zitiao Network Technology Co Ltd
Current assignee: Beijing Zitiao Network Technology Co Ltd
Priority date: 2021-12-31
Filing date: 2021-12-31
Publication date: 2022-04-12

Abstract

The disclosure relates to an object classification method, an object classification device, a readable medium and an electronic device. The method comprises the following steps: obtaining a plurality of first object features of an object to be classified; inputting the first object characteristics into a target classification model to obtain a target class corresponding to the object to be classified; the target classification model is a classification model obtained by training a preset rule classification model through a second object feature, and the second object feature is a feature obtained by performing feature simplification processing on the plurality of first object features. Therefore, the object features are simplified and then trained and recognized, so that the computation amount of the model can be reduced, and the model training efficiency and the object classification efficiency can be improved.

Description

Object classification method and device, readable medium and electronic equipment

Technical Field

The present disclosure relates to the field of computer technologies, and in particular, to an object classification method, an object classification device, a readable medium, and an electronic device.

Background

With the development of internet technology, the internet has appeared with more traffic type products, such as video APP (Application), and these traffic type products are prone to black production problems, such as that some black production users may perform illegal advertisement pushing, illegal traffic brushing, and illegal Application pushing, etc. In order to identify the black users, manual inspection identification can be performed in the related art, but manual inspection needs a lot of manpower, and the efficiency is low.

Disclosure of Invention

This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

In a first aspect, the present disclosure provides an object classification method, including:

obtaining a plurality of first object features of an object to be classified;

inputting the first object characteristics into a target classification model to obtain a target class corresponding to the object to be classified;

the target classification model is obtained by training a preset rule classification model through a second object feature, and the second object feature is obtained by performing feature simplification processing on the plurality of first object features.

In a second aspect, the present disclosure provides an object classification apparatus, the apparatus comprising:

the characteristic acquisition module is used for acquiring a plurality of first object characteristics of the object to be classified;

the object classification module is used for inputting the first object characteristics into a target classification model to obtain a target class corresponding to the object to be classified;

In a third aspect, the present disclosure provides a computer readable medium having stored thereon a computer program which, when executed by a processing apparatus, performs the steps of the method of the first aspect of the present disclosure.

In a fourth aspect, the present disclosure provides an electronic device comprising:

a storage device having a computer program stored thereon;

processing means for executing the computer program in the storage means to implement the steps of the method of the first aspect of the present disclosure.

By adopting the technical scheme, a plurality of first object characteristics of the object to be classified are obtained; inputting the first object characteristics into a target classification model to obtain a target class corresponding to the object to be classified; the target classification model is a classification model obtained by training a preset rule classification model through a second object feature, and the second object feature is a feature obtained by performing feature simplification processing on the plurality of first object features. Therefore, the object features are simplified and then trained and recognized, so that the computation amount of the model can be reduced, and the model training efficiency and the object classification efficiency can be improved.

Additional features and advantages of the disclosure will be set forth in the detailed description which follows.

Drawings

The above and other features, advantages and aspects of various embodiments of the present disclosure will become more apparent by referring to the following detailed description when taken in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numbers refer to the same or similar elements. It should be understood that the drawings are schematic and that elements and features are not necessarily drawn to scale. In the drawings:

FIG. 1 is a flow diagram illustrating a method of object classification according to an exemplary embodiment.

FIG. 2 is a flow diagram illustrating a method of training a target classification model according to an exemplary embodiment.

Fig. 3 is a block diagram illustrating an object classification apparatus according to an exemplary embodiment.

Fig. 4 is a block diagram illustrating another object classification apparatus according to an example embodiment.

FIG. 5 is a block diagram illustrating an electronic device in accordance with an example embodiment.

Detailed Description

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it is to be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but rather are provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the disclosure are for illustration purposes only and are not intended to limit the scope of the disclosure.

It should be understood that the various steps recited in the method embodiments of the present disclosure may be performed in a different order, and/or performed in parallel. Moreover, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present disclosure is not limited in this respect.

The term "include" and variations thereof as used herein are open-ended, i.e., "including but not limited to". The term "based on" is "based, at least in part, on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Relevant definitions for other terms will be given in the following description.

It should be noted that the terms "first", "second", and the like in the present disclosure are only used for distinguishing different devices, modules or units, and are not used for limiting the order or interdependence relationship of the functions performed by the devices, modules or units.

It is noted that references to "a", "an", and "the" modifications in this disclosure are intended to be illustrative rather than limiting, and that those skilled in the art will recognize that "one or more" may be used unless the context clearly dictates otherwise.

The names of messages or information exchanged between devices in the embodiments of the present disclosure are for illustrative purposes only, and are not intended to limit the scope of the messages or information.

First, an application scenario of the present disclosure will be explained. The present disclosure may be applied to object classification scenarios, in particular scenarios where objects (e.g. users, devices, accounts or specific data, etc.) in the internet are classified in order to distinguish between risky objects and non-risky objects. In internet applications such as video APP, some black products users can perform illegal advertisement pushing, illegal traffic brushing, illegal application pushing and the like, in order to identify the black products users, manual inspection identification can be performed in the related technology, but manual inspection needs a large amount of manpower, and efficiency is low.

In addition, the features of various modalities can be fused through a classification model (for example, a model such as XGBoost), but in actual use, the original features of an input object are more, so that the model is complex, and the efficiency in the model training process and the model application process is low.

In order to solve the above problems, the present disclosure provides an object classification method, an apparatus, a readable medium, and an electronic device, which may simplify a first object feature of a sample object to obtain a second object feature in a training process, train a preset rule classification model according to the simplified second object feature and the sample object to obtain a target classification model, and classify an object to be classified by the target classification model to obtain a target class. Therefore, the calculation amount of the model can be reduced by simplifying the characteristics, and the model training efficiency and the object classification efficiency are improved.

The following detailed description of specific embodiments of the present disclosure is provided in connection with the accompanying drawings.

Fig. 1 illustrates an object classification method according to an exemplary embodiment, as shown in fig. 1, the method including:

step 101, obtaining a plurality of first object features of an object to be classified.

And 102, inputting the first object characteristics into a target classification model to obtain a target class corresponding to the object to be classified.

The first object feature may be an original feature of the object to be classified, which may be input by a user, or may be extracted from the object to be classified by a feature extraction method. By way of example, the first object feature may comprise: the method comprises the steps of terminal type, terminal operating system version, account category, account online time, account attention and the like.

The above-mentioned manner of performing the feature reduction processing on the plurality of first object features may include: deleting low value features or deleting duplicate features. The target classification model can process the second object features according to the second object features, so that redundant feature processing can be reduced, the model operation amount can be reduced, and the object classification efficiency can be improved.

By adopting the method, a plurality of first object characteristics of the object to be classified are obtained; inputting the first object characteristics into a target classification model to obtain a target class corresponding to the object to be classified; the target classification model is a classification model obtained by training a preset rule classification model through a second object feature, and the second object feature is a feature obtained by performing feature simplification processing on the plurality of first object features. Therefore, the object features are simplified and then trained and recognized, so that the computation amount of the model can be reduced, and the model training efficiency and the object classification efficiency can be improved.

FIG. 2 is a flowchart illustrating a method of training a target classification model according to an exemplary embodiment, which may include, as shown in FIG. 2:

step 201, a plurality of sample objects are obtained.

Wherein each sample object comprises a plurality of first object features.

The sample object may be, for example, a user, a device, an account number, or specific data, etc. The first object feature may be an original feature of the sample object, may be input by a user, and may also be extracted from the object to be classified by means of feature extraction. By way of example, the first object feature may comprise: the method comprises the steps of terminal type, terminal operating system version, account category, account online time, account attention and the like.

Step 202, performing feature reduction processing on the plurality of first object features to determine second object features.

In this step, the feature simplification process may include any one or more of the following:

and a first characteristic simplifying mode, namely simplifying the first characteristic according to the characteristic value so as to screen out part of low-value characteristics.

Illustratively, the first mode may include the following steps: firstly, acquiring the characteristic value of each first object characteristic, and taking the first object characteristic with the characteristic value within a preset value range as a first candidate characteristic; then, the second object feature is obtained according to the first candidate feature.

Wherein the feature value may include a feature uniqueness degree and/or an information value, wherein the feature uniqueness degree may be calculated according to the number of the sample objects and the number of non-repetition values of the first object feature; the information value may be used to characterize the relevance of the first object feature to the object classification result.

In one embodiment of the present disclosure, the feature value may include a feature uniqueness, and the feature uniqueness may be calculated according to the following formula (1):

Uniqueness＝feature.nunique/len(samples) (1)

wherein uniquess represents the feature Uniqueness of the first object feature; len (samples) represents the number of sample objects; nunique denotes the number of non-repetitive values of the first object feature.

For example, the total number of the sample objects is 100, each sample object has 3 first object features, wherein if there are 2 non-repetitive values in 100 sample objects for a first specific first object feature, the feature uniqueness of the first specific first object feature is 2/100, that is, 0.02; if there are 50 non-repetitive values in 100 sample objects for a second particular first object feature, the feature uniqueness of this second particular first object feature is 50/100, i.e. 0.5; if there are 90 non-repetitive values of the third specific first object feature in 100 sample objects, the feature uniqueness of the third specific first object feature is 90/100, i.e. 0.9.

Further, in the case that the feature value includes the feature uniqueness, the preset value range may include a range from a first threshold to a second threshold, for example, a range greater than or equal to the first threshold and less than or equal to the second threshold, or a range greater than or equal to the first threshold and less than or equal to the second threshold. For example, the first threshold may be 0.02, the second threshold may be 0.7, and in the above example, the feature uniqueness of the second specific first object feature is within the preset value range, and the feature uniqueness of the first specific first object feature and the third specific first object feature is not within the preset value range, so that the second specific first object feature may be used as the simplified first candidate feature.

Further, the first threshold may be determined according to the number of the sample objects, and for example, the first threshold may be obtained by setting a first preset value at the sample objects, where the first preset value may be 1 or 2.

It should be noted that the first object features with too large or too small feature uniqueness do not have effective information required by object classification, which wastes training resources. Furthermore, for first object features that are too unique, participation in training can also cause problems with model overfitting.

Therefore, by the mode, the first object features with the characteristic uniqueness which is too large or too small are screened out, and the first object features with the characteristic uniqueness which is in the preset value range are reserved, so that the training complexity can be reduced, the problem of model overfitting is reduced, and the model training efficiency and the accuracy of the model are improved.

In another embodiment of the present disclosure, the feature Value may include an Information Value (IV), which may be used to characterize the correlation of the first object feature with the object classification result, for example, the higher the Information Value of a specific first object feature is, the higher the correlation of the specific first object feature with the object classification result is. The information value calculation method can refer to the mode in the related technology, and details are not repeated in the disclosure.

Further, in the case where the characteristic value includes an information value, the preset value range may include a range greater than or equal to a third threshold value, that is, the information value greater than or equal to the third threshold value is within the preset value range, and the third threshold value may be 0.01 or 0.02, for example.

Therefore, by the mode, the first object features with the too small information value are screened out, and the first object features with the relatively large information value are reserved, so that the training complexity can be reduced, and the model training efficiency is improved.

In another embodiment of the present disclosure, the feature value may include a feature uniqueness degree and an information value, so that the first object feature having the feature uniqueness degree within a first preset value range may be first screened according to the feature uniqueness degree as a first predetermined feature; and then screening the first to-be-determined feature according to the information value, taking the first to-be-determined feature with the information value within a second preset value range as a second to-be-determined feature, and taking the second to-be-determined feature as the first candidate feature.

Therefore, because the calculation amount of the feature uniqueness degree is small and the calculation amount of the information value is large, by the method, a part of first object features can be screened out through the small calculation amount, and then the rest first to-be-determined features are screened out according to the information value, so that the calculation efficiency can be further improved.

Further, the first candidate feature may be directly used as the second object feature; the first candidate feature may also be further filtered to obtain a second object feature.

For example, in the case that there are a plurality of first candidate features, the screening may be performed according to the relevance of the plurality of first candidate features, and the method may include the following steps:

firstly, obtaining the correlation between any two first candidate characteristics; and taking the first candidate feature with the correlation degree larger than or equal to a preset correlation degree threshold value as a repeated candidate feature.

Next, a feature having the largest information value among the repeated candidate features is set as a second candidate feature.

The correlation may be, for example, a correlation coefficient, such as a sparman correlation coefficient, a Kendall correlation coefficient, or a Pearson correlation coefficient. The preset correlation threshold may be 0.7 or 0.8. If the correlation degree of the two first candidate features is 0.9, the feature with the largest information value in the two first candidate features can be used as a second candidate feature; if the information values of the two first candidate features are equal, one of the first candidate features can be randomly selected as the second candidate feature.

Therefore, repeated first candidate features can be screened out according to the similarity, model training and model application are further simplified, and therefore model training efficiency can be improved, and the classification effect of the model cannot be influenced.

And finally, acquiring the second object characteristic according to the second candidate characteristic.

For example, the second candidate feature may be the second object feature; and performing chi-square binning processing according to the second candidate feature to obtain a second object feature.

It should be noted that the sample object further includes a classification label, and the classification label can be used as a supervision signal of the sample object for the chi-square binning process. The card square binning processing may refer to methods in the related art, which are not described in detail in this disclosure.

Step 203, training the preset rule classification model according to the plurality of sample objects and the second object characteristics to obtain the target classification model.

In the above manner, a plurality of sample objects may be obtained, each of which includes a plurality of first object features; carrying out feature simplification processing on the plurality of first object features to determine second object features; and training the preset rule classification model according to the plurality of sample objects and the second object characteristics to obtain the target classification model. In this way, the object features can be simplified, thereby improving the efficiency of model training.

In another embodiment of the present disclosure, the first object feature may include a continuous feature and/or a discrete feature, the discrete feature is used for characterizing feature values as discrete data, and the continuous feature may be used for characterizing feature values as continuous values, so that the feature reduction processing is performed on a plurality of first object features in the step 202, and determining the second object feature may include the following steps:

firstly, carrying out logarithmic rounding on the continuous features according to a preset classification scale to obtain third candidate features.

Illustratively, the continuous features may be rounded logarithmically by the following equation (2):

approx_feature＝exp(round(log(feature+a),digit)) (2)

wherein the approx _ feature represents a third candidate feature; feature represents a continuous feature, which is any number greater than or equal to 0; a represents a preset feature bias, which may be a small value to avoid the continuous feature being 0, e.g., 1 e-5; digit represents a preset classification scale that may be used to characterize the number of decimal places for logarithmic rounding, e.g., digit is 1, representing rounding to the first digit after the decimal point; log (feature + a) represents taking the logarithm of feature + a; round (log (feature + a), digit) means rounding log (feature + a) to the second digit; exp (round (log (feature + a), digit)) means that round (log (feature + a), digit) is exponentially reduced.

It should be noted that, since the principle of card-direction binning is a process of gradually merging from bottom to top, if the original features are too dispersed, the binning speed is very slow. For this reason, some approximate rounding of the original continuous features is required so that the subsequent chi-squared binning step can be accelerated. Since most of the service sensitivity of continuous features approximately presents a logarithmic scale, the logarithmic rounding can be carried out, so that the original information can be kept as much as possible while the service sensitivity scale is considered.

Further, in this step, logarithmic rounding may be performed for consecutive features having a feature uniqueness greater than or equal to a preset uniqueness threshold; and for feature uniqueness less than the preset uniqueness threshold, logarithmic rounding may not be performed.

And then, acquiring the second object characteristic according to the third candidate characteristic.

In this step, chi-square binning processing may be performed on the third candidate features according to the classification labels and a preset binning number to obtain fourth candidate features; then, the fourth candidate feature is taken as the second object feature.

The preset binning number is used for representing the binning number of the fourth candidate feature after chi-square binning processing.

Further, in the above-mentioned card square binning processing on the third candidate feature, a binning configuration parameter of a fourth candidate feature after the card square binning processing may also be obtained, where the binning configuration parameter includes a corresponding relationship between a value of the fourth candidate feature and a value of the first object feature.

It should be noted that, in the case that the first object feature is a continuous feature, the corresponding relationship between the value of the fourth candidate feature and the value of the first object feature may include an upper limit and a lower limit of each box.

For example, if the third candidate feature is an arbitrary number value where the continuous feature is greater than or equal to 0 and the preset binning number is 3, after the chi-square binning processing is performed on the third candidate feature according to the classification label and the preset binning number, the third candidate feature may be divided into three sections, namely, greater than or equal to 0 and less than 40, greater than or equal to 40 and less than 60, and the fourth candidate feature corresponding to each section may be a box number 1, 2, or 3, where the lower limit of the binning configuration parameter corresponding to the fourth candidate feature 1 is 0 and the upper limit is 40; the lower limit and the upper limit of the box configuration parameter corresponding to the fourth candidate feature 2 are 40 and 60 respectively; the lower limit of the box configuration parameter corresponding to the fourth candidate feature 1 is 60, and the upper limit is infinity; in this way, the continuous third candidate feature may be converted into a discrete fourth candidate feature, and the binning configuration parameters may be obtained to use the fourth candidate feature as the second object feature.

Further, in a case that the first object feature is a discrete feature, the correspondence between the values of the fourth candidate features and the values of the first object feature may include the value of the first object feature corresponding to each value of the fourth candidate feature. Illustratively, the first object feature is a city, the fourth candidate feature is a city type obtained by classifying the city, values of the city type include a1, a2 and A3, values of the city include C1, C2, C3, C4, C5 and C6, a value of the city corresponding to a city type a1 is C1, a value of the city corresponding to a city type a2 includes C2, C3 and C4, and a value of the city corresponding to a city type A3 includes C5 and C6, so that a corresponding relationship between the value of the fourth candidate feature and the value of the first object feature, that is, the binning configuration parameter, can be obtained.

Then, in step 203, training the preset rule classification model according to the plurality of sample objects and the second object feature to obtain the target classification model, which may include the following steps:

firstly, training the preset rule classification model according to the plurality of sample objects and the second object characteristics to obtain a to-be-classified model.

The predetermined Rule classification model may include any Rule table model, for example, the predetermined Rule classification model may include an SBRL (Scalable Bayesian Rule list), and in this step, a training task of the SBRL may be executed according to a plurality of sample objects and second object features to obtain a to-be-determined classification model.

And then, updating the classification rule in the to-be-classified model according to the box-dividing configuration parameters to obtain a target classification model.

For example, if the classification rule in the to-be-classified model is classified according to the box serial number of the fourth candidate feature, the box serial number of the fourth candidate feature may be converted into a binning configuration parameter corresponding to the fourth candidate feature according to the analysis configuration parameter, so that the binning configuration parameter is updated to the classification rule according to the binning configuration parameter, thereby obtaining the target classification model.

Therefore, when the model is applied, the judgment of the classification rule can be carried out directly according to the continuous features in the first object features of the object to be classified without carrying out feature binning again, and the object to be classified is efficiently classified. Meanwhile, after the classification rules in the classification model to be determined are updated according to the box configuration parameters, the readability of the classification rules of the target classification model is improved, and the target classification model is convenient for a user to use.

Fig. 3 is a block diagram illustrating an object classification apparatus according to an exemplary embodiment. As shown in fig. 3, the object classification apparatus includes:

a feature obtaining module 301, configured to obtain a plurality of first object features of an object to be classified;

an object classification module 302, configured to input the multiple first object features into a target classification model, so as to obtain a target class corresponding to the object to be classified;

Fig. 4 is a block diagram illustrating another object classification apparatus according to an example embodiment. As shown in fig. 4, the object classification apparatus further includes:

a model training module 401 for obtaining a plurality of sample objects, each of the sample objects comprising a plurality of first object features; carrying out feature simplification processing on the plurality of first object features to determine second object features; and training the preset rule classification model according to the plurality of sample objects and the second object characteristics to obtain the target classification model.

Optionally, the model training module 401 is configured to obtain a feature value of each first object feature; taking the first object feature with the feature value within a preset value range as the first candidate feature; and acquiring the second object characteristic according to the first candidate characteristic.

Optionally, the feature value comprises a feature uniqueness calculated from the number of sample objects and the number of non-repetitive values of the first object feature and/or an information value characterizing a relevance of the first object feature to the object classification result.

Optionally, the first object feature includes a continuous feature, the continuous feature is used to characterize a feature with a feature value of a continuous numerical value, and the model training module 401 is configured to perform logarithmic rounding on the continuous feature according to a preset classification scale to obtain a third candidate feature; and acquiring the second object characteristic according to the third candidate characteristic.

Optionally, the sample object further comprises a classification label; the model training module 401 is configured to perform chi-square binning processing on the third candidate feature according to the classification label and a preset binning number to obtain a fourth candidate feature; and taking the fourth candidate feature as the second object feature.

Optionally, the model training module 401 is further configured to obtain a binning configuration parameter of the fourth candidate feature after chi-square binning processing; the box configuration parameters comprise the corresponding relation between the value of the fourth candidate feature and the value of the first object feature; training the preset rule classification model according to the plurality of sample objects and the second object characteristics to obtain a to-be-classified model; and updating the classification rule in the classification model to be determined according to the box-dividing configuration parameters to obtain a target classification model.

With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

Referring now to fig. 5, a schematic diagram of an electronic device (e.g., a terminal device or server) 900 suitable for use in implementing embodiments of the present disclosure is shown. The terminal device in the embodiments of the present disclosure may include, but is not limited to, a mobile terminal such as a mobile phone, a notebook computer, a digital broadcast receiver, a PDA (personal digital assistant), a PAD (tablet computer), a PMP (portable multimedia player), a vehicle terminal (e.g., a car navigation terminal), and the like, and a stationary terminal such as a digital TV, a desktop computer, and the like. The electronic device shown in fig. 5 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.

As shown in fig. 5, the electronic device 900 may include a processing means (e.g., a central processing unit, a graphics processor, etc.) 901 that may perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)902 or a program loaded from a storage means 908 into a Random Access Memory (RAM) 903. In the RAM903, various programs and data necessary for the operation of the electronic apparatus 900 are also stored. The processing apparatus 901, the ROM902, and the RAM903 are connected to each other through a bus 904. An input/output (I/O) interface 905 is also connected to bus 904.

Generally, the following devices may be connected to the I/O interface 905: input devices 906 including, for example, a touch screen, touch pad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; an output device 907 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 908 including, for example, magnetic tape, hard disk, etc.; and a communication device 909. The communication device 909 may allow the electronic apparatus 900 to perform wireless or wired communication with other apparatuses to exchange data. While fig. 5 illustrates an electronic device 900 having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided.

In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a non-transitory computer readable medium, the computer program containing program code for performing the method illustrated by the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication device 909, or installed from the storage device 908, or installed from the ROM 902. The computer program performs the above-described functions defined in the methods of the embodiments of the present disclosure when executed by the processing apparatus 901.

It should be noted that the computer readable medium in the present disclosure can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In contrast, in the present disclosure, a computer readable signal medium may comprise a propagated data signal with computer readable program code embodied therein, either in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.

In some embodiments, the clients, servers may communicate using any currently known or future developed network Protocol, such as HTTP (HyperText Transfer Protocol), and may interconnect with any form or medium of digital data communication (e.g., a communications network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), the Internet (e.g., the Internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed network.

The computer readable medium may be embodied in the electronic device; or may exist separately without being assembled into the electronic device.

The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: obtaining a plurality of first object features of an object to be classified; inputting the first object characteristics into a target classification model to obtain a target class corresponding to the object to be classified; the target classification model is obtained by training a preset rule classification model through a second object feature, and the second object feature is obtained by performing feature simplification processing on the plurality of first object features.

Computer program code for carrying out operations for the present disclosure may be written in any combination of one or more programming languages, including but not limited to an object oriented programming language such as Java, Smalltalk, C + +, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The modules described in the embodiments of the present disclosure may be implemented by software or hardware. The name of a module does not in some cases constitute a definition of the module itself, and for example, the feature acquisition module may also be described as a "module that acquires a plurality of first object features of an object to be classified".

The functions described herein above may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), systems on a chip (SOCs), Complex Programmable Logic Devices (CPLDs), and the like.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

Example 1 provides an object classification method according to one or more embodiments of the present disclosure, including:

obtaining a plurality of first object features of an object to be classified;

Example 2 provides the method of example 1, the target classification model being trained by:

obtaining a plurality of sample objects, each of the sample objects comprising a plurality of first object features;

performing feature simplification processing on the plurality of first object features to determine second object features;

and training the preset rule classification model according to the plurality of sample objects and the second object characteristics to obtain the target classification model.

Example 3 provides the method of example 2, wherein performing feature reduction processing on the plurality of first object features and determining second object features comprises:

acquiring the feature value of each first object feature;

taking the first object feature with the feature value within a preset value range as the first candidate feature;

and acquiring the second object characteristic according to the first candidate characteristic.

Example 4 provides the method of example 3, the feature value including a feature uniqueness calculated from the number of sample objects and the non-repetition value of the first object feature and/or an information value characterizing a relevance of the first object feature to an object classification result, according to one or more embodiments of the present disclosure.

Example 5 provides the method of example 2, the first object feature includes a continuous feature for characterizing features whose feature values are continuous numerical values, the performing feature reduction processing on the plurality of first object features, and determining the second object feature includes:

carrying out logarithmic rounding on the continuous features according to a preset classification scale to obtain third candidate features;

and acquiring the second object characteristic according to the third candidate characteristic.

Example 6 provides the method of example 5, the sample object further comprising a classification label, in accordance with one or more embodiments of the present disclosure; the obtaining the second object feature according to the third candidate feature includes:

according to the classification labels and the preset box dividing number, carrying out chi-square box dividing processing on the third candidate features to obtain fourth candidate features;

and taking the fourth candidate feature as the second object feature.

Example 7 provides the method of example 6, further comprising, in accordance with one or more embodiments of the present disclosure:

acquiring a binning configuration parameter of the fourth candidate feature after chi-square binning processing; the box-dividing configuration parameters comprise the corresponding relation between the value of the fourth candidate characteristic and the value of the first object characteristic;

the training the preset rule classification model according to the plurality of sample objects and the second object characteristics to obtain the target classification model comprises:

training the preset rule classification model according to the plurality of sample objects and the second object characteristics to obtain a to-be-classified model;

and updating the classification rules in the to-be-classified model according to the box-dividing configuration parameters to obtain a target classification model.

Example 8 provides, in accordance with one or more embodiments of the present disclosure, an object classification apparatus, the apparatus comprising:

Example 9 provides the apparatus of example 8, the apparatus further comprising, in accordance with one or more embodiments of the present disclosure:

a model training module for obtaining a plurality of sample objects, each of the sample objects comprising a plurality of first object features; carrying out feature simplification processing on the plurality of first object features to determine second object features; and training the preset rule classification model according to the plurality of sample objects and the second object characteristics to obtain the target classification model.

Example 10 provides the apparatus of example 9, the model training module to obtain a feature value for each first object feature; taking the first object feature with the feature value within a preset value range as the first candidate feature; and acquiring the second object characteristic according to the first candidate characteristic.

Example 11 provides the apparatus of example 10, the feature value including a feature uniqueness calculated from the number of sample objects and the non-repetition value of the first object feature and/or an information value characterizing a relevance of the first object feature to an object classification result, according to one or more embodiments of the present disclosure.

Example 12 provides the apparatus of example 9, where the first object feature includes a continuous feature, and the continuous feature is used to characterize a feature with a feature value of a continuous numerical value, and the model training module is used to perform logarithmic rounding on the continuous feature according to a preset classification scale to obtain a third candidate feature; and acquiring the second object characteristic according to the third candidate characteristic.

Example 13 provides the apparatus of example 12, the sample object further comprising a classification label, in accordance with one or more embodiments of the present disclosure; the model training module 401 is configured to perform chi-square binning processing on the third candidate features according to the classification labels and a preset binning number to obtain fourth candidate features; and taking the fourth candidate feature as the second object feature.

Example 14 provides the apparatus of example 13, the model training module further configured to obtain binning configuration parameters of the fourth candidate feature after chi-squared binning processing; the box-dividing configuration parameters comprise the corresponding relation between the value of the fourth candidate characteristic and the value of the first object characteristic; training the preset rule classification model according to the plurality of sample objects and the second object characteristics to obtain a to-be-classified model; and updating the classification rules in the to-be-classified model according to the box-dividing configuration parameters to obtain a target classification model.

The foregoing description is only exemplary of the preferred embodiments of the disclosure and is illustrative of the principles of the technology employed. It will be appreciated by those skilled in the art that the scope of the disclosure herein is not limited to the particular combination of features described above, but also encompasses other embodiments in which any combination of the features described above or their equivalents does not depart from the spirit of the disclosure. For example, the above features and (but not limited to) the features disclosed in this disclosure having similar functions are replaced with each other to form the technical solution.

Further, while operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while several specific implementation details are included in the above discussion, these should not be construed as limitations on the scope of the disclosure. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination.

Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims. With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

Claims

1. A method of object classification, the method comprising:

obtaining a plurality of first object features of an object to be classified;

2. The method of claim 1, wherein the object classification model is trained by:

3. The method of claim 2, wherein the performing a feature reduction process on the plurality of first object features and determining second object features comprises:

acquiring the feature value of each first object feature;

4. The method of claim 3, wherein the feature value comprises a feature uniqueness calculated from the number of sample objects and the number of non-repetitive values of the first object feature and/or an information value characterizing a relevance of the first object feature to an object classification result.

5. The method of claim 2, wherein the first object features comprise continuous features for characterizing features having feature values that are continuous values, and wherein performing feature reduction processing on the plurality of first object features and determining second object features comprises:

6. The method of claim 5, wherein the sample object further comprises a classification label; the obtaining the second object feature according to the third candidate feature includes:

and taking the fourth candidate feature as the second object feature.

7. The method of claim 6, further comprising:

8. An object classification apparatus, characterized in that the apparatus comprises:

9. A computer-readable medium, on which a computer program is stored, characterized in that the program, when being executed by processing means, carries out the steps of the method of any one of claims 1 to 7.

10. An electronic device, comprising:

a storage device having a computer program stored thereon;

processing means for executing the computer program in the storage means to carry out the steps of the method according to any one of claims 1 to 7.