CN113449923A - Multi-model object market quotation prediction method and device - Google Patents

Multi-model object market quotation prediction method and device Download PDF

Info

Publication number
CN113449923A
CN113449923A CN202110780081.1A CN202110780081A CN113449923A CN 113449923 A CN113449923 A CN 113449923A CN 202110780081 A CN202110780081 A CN 202110780081A CN 113449923 A CN113449923 A CN 113449923A
Authority
CN
China
Prior art keywords
data
market
information
user
prediction
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110780081.1A
Other languages
Chinese (zh)
Inventor
龚孟旭
陈冰
于娇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Bank of China Ltd
Original Assignee
Bank of China Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Bank of China Ltd filed Critical Bank of China Ltd
Priority to CN202110780081.1A priority Critical patent/CN113449923A/en
Publication of CN113449923A publication Critical patent/CN113449923A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/067—Enterprise or organisation modelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Theoretical Computer Science (AREA)
  • Development Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • General Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Tourism & Hospitality (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Game Theory and Decision Science (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Technology Law (AREA)
  • Educational Administration (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention provides a multi-model object market forecasting method and a multi-model object market forecasting device, which belong to artificial intelligence, and comprise the following steps: acquiring historical data of a target object; preprocessing historical data of the object to determine preprocessed data; extracting the characteristics of the preprocessed data, and establishing a prediction model; dividing the preprocessed data into a training set and a testing set, and determining a market condition prediction model based on technical indexes, a market condition prediction model based on information and a transaction emotion model based on user guessing fluctuation by using the training set; respectively determining the accuracy of the respective prediction results of the three models according to the test set; according to the accuracy of the prediction results of the three models, weights are distributed to the market forecasting model based on the technical indexes, the market forecasting model based on the information and the trading emotion model based on the guessing of the fluctuation of the user, the market of the current object is predicted after weighted average, and the market forecasting result of the current object is determined. The method effectively improves the prediction accuracy of the target object.

Description

Multi-model object market quotation prediction method and device
Technical Field
The invention relates to the technical field of computer data processing, in particular to a multi-model object market quotation prediction method and device.
Background
This section is intended to provide a background or context to the embodiments of the invention that are recited in the claims. The description herein is not admitted to be prior art by inclusion in this section.
In recent years, with the development of big data, distributed and artificial intelligence technologies, enterprises related to the big financial industry have been researched and researched for forecasting the market quotation of the subject matter in the technical field.
At present, in the prior art, the market condition trend of a target is predicted by mainly collecting historical technical index information of the target and carrying out machine learning modeling analysis, and because the market condition trend of the target has great uncertainty, the market condition trend is influenced by objective factors such as the historical trend and technical indexes of the target and main factors such as public opinions and human emotions, the complexity of modeling and predicting the market condition by singly depending on the historical trend and the technical index data of the target and being difficult to restore the market condition prediction is influenced, and the accuracy of prediction is influenced.
Therefore, how to provide a new solution, which can solve the above technical problems, is a technical problem to be solved in the art.
Disclosure of Invention
The embodiment of the invention provides a multi-model object market quotation prediction method, which effectively improves the prediction accuracy of an object, and comprises the following steps:
acquiring historical data of a target object;
preprocessing historical data of the object to determine preprocessed data;
extracting the characteristics of the preprocessed data, and establishing a prediction model;
dividing the preprocessed data into a training set and a testing set, and determining a market condition prediction model based on technical indexes, a market condition prediction model based on information and a transaction emotion model based on user guessing fluctuation by using the training set;
respectively determining the accuracy rates of the prediction results of a market forecasting model based on technical indexes, a market forecasting model based on information and a user guessing fluctuation transaction emotion model according to the test set;
according to the accuracy of the prediction results of the three models, weights are distributed to the market forecasting model based on the technical indexes, the market forecasting model based on the information and the trading emotion model based on the guessing of the fluctuation of the user, the market of the current object is predicted after weighted average, and the market forecasting result of the current object is determined.
The embodiment of the invention also provides a multi-model object market prediction device, which comprises:
the system comprises a target historical data acquisition module, a data processing module and a data processing module, wherein the target historical data acquisition module is used for acquiring target historical data;
the preprocessing module is used for preprocessing the historical data of the object and determining preprocessed data;
the characteristic extraction module is used for extracting the characteristics of the preprocessed data and establishing a prediction model;
the model training module is used for dividing the preprocessed data into a training set and a testing set, and determining a market forecasting model based on technical indexes, a market forecasting model based on information and a user guessing fluctuation transaction emotion model by using the training set;
the prediction result accuracy rate determining module is used for respectively determining the respective prediction result accuracy rates of the market condition prediction model based on the technical indexes, the market condition prediction model based on the information and the user guessing fluctuation transaction emotion model according to the test set;
and the current target object market forecasting module is used for distributing weights to the market forecasting model based on the technical indexes, the market forecasting model based on the information and the user guessing fluctuation transaction emotion model according to the respective forecasting result accuracy of the three models, forecasting the current target object market after weighted averaging, and determining the current target object market forecasting result.
The embodiment of the invention also provides computer equipment which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein when the processor executes the computer program, the multi-model object market situation prediction method is realized.
An embodiment of the present invention further provides a computer-readable storage medium, where a computer program for executing the multi-model object market prediction method is stored in the computer-readable storage medium.
The embodiment of the invention provides a method and a device for predicting the market quotation of a multi-model object, which comprises the following steps: firstly, acquiring historical data of a target object; then, preprocessing the historical data of the object to determine preprocessed data; then, extracting the characteristics of the preprocessed data, and establishing a prediction model; continuously dividing the preprocessed data into a training set and a testing set, and determining a market condition prediction model based on technical indexes, a market condition prediction model based on information and a transaction emotion model based on user guessing fluctuation by using the training set; next, respectively determining the accuracy of the prediction results of a market forecasting model based on technical indexes, a market forecasting model based on information and a user guessing fluctuation transaction emotion model according to the test set; and finally, according to the accuracy of the prediction results of the three models, distributing weights to the market quotation prediction model based on the technical indexes, the market quotation prediction model based on the information and the user guessing fluctuation transaction emotion model, predicting the market quotation of the current target after weighted averaging, and determining the market quotation prediction result of the current target. The method comprises the steps of carrying out data preprocessing, characteristic engineering and other operations on historical data of different types of objects, building models, providing market intelligent prediction for users from multiple dimensions, not only comprising traditional market prediction in a historical technical index mode, but also comprising market prediction based on information and transaction emotion analysis of the users, training to obtain a market prediction model based on technical indexes, a market prediction model based on information and a transaction emotion model based on user guessing, completing object market prediction by carrying out weight distribution on the three models, and obtaining a final prediction result by different weight weighted averages of different distribution weights of prediction accuracy rates of the three models, so that the prediction accuracy rate of the objects is effectively improved. The invention not only applies the object history technical index as the data acquisition object, but also brings the real-time information and the user emotion into the data acquisition category, thereby increasing the credibility and the accuracy of the model. According to the invention, the market situation intelligent prediction is provided for the user from multiple dimensions, more target object investment prediction information can be provided for the user, the user can know market situation analysis of the product from multiple directions and multiple dimensions, the user experience and the user viscosity are effectively improved, the user is further attracted, and the possibility is provided for realizing deep conversion of the user.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts. In the drawings:
fig. 1 is a schematic diagram of a multi-model object market prediction method according to an embodiment of the present invention.
Fig. 2 is a schematic process diagram of a multi-model object market forecasting method for determining a market forecasting model based on technical indexes according to an embodiment of the present invention.
Fig. 3 is a schematic process diagram of determining an information-based market prediction model according to a multi-model object market prediction method of the present invention.
Fig. 4 is a schematic diagram of a computer device for executing a multi-model object market prediction method implemented by the present invention.
Fig. 5 is a schematic diagram of a multi-model object market prediction device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.
The invention belongs to artificial intelligence. Fig. 1 is a schematic diagram of a multi-model object market prediction method according to an embodiment of the present invention, and as shown in fig. 1, an embodiment of the present invention provides a multi-model object market prediction method, which effectively improves the prediction accuracy of an object, and includes:
step 101: acquiring historical data of a target object;
step 102: preprocessing historical data of the object to determine preprocessed data;
step 103: extracting the characteristics of the preprocessed data, and establishing a prediction model;
step 104: dividing the preprocessed data into a training set and a testing set, and determining a market condition prediction model based on technical indexes, a market condition prediction model based on information and a transaction emotion model based on user guessing fluctuation by using the training set;
step 105: respectively determining the accuracy rates of the prediction results of a market forecasting model based on technical indexes, a market forecasting model based on information and a user guessing fluctuation transaction emotion model according to the test set;
step 106: according to the accuracy of the prediction results of the three models, weights are distributed to the market forecasting model based on the technical indexes, the market forecasting model based on the information and the trading emotion model based on the guessing of the fluctuation of the user, the market of the current object is predicted after weighted average, and the market forecasting result of the current object is determined.
The embodiment of the invention provides a multi-model object market quotation prediction method, which comprises the following steps: firstly, acquiring historical data of a target object; then, preprocessing the historical data of the object to determine preprocessed data; then, extracting the characteristics of the preprocessed data, and establishing a prediction model; continuously dividing the preprocessed data into a training set and a testing set, and determining a market condition prediction model based on technical indexes, a market condition prediction model based on information and a transaction emotion model based on user guessing fluctuation by using the training set; next, respectively determining the accuracy of the prediction results of a market forecasting model based on technical indexes, a market forecasting model based on information and a user guessing fluctuation transaction emotion model according to the test set; and finally, according to the accuracy of the prediction results of the three models, distributing weights to the market quotation prediction model based on the technical indexes, the market quotation prediction model based on the information and the user guessing fluctuation transaction emotion model, predicting the market quotation of the current target after weighted averaging, and determining the market quotation prediction result of the current target. The method comprises the steps of carrying out data preprocessing, characteristic engineering and other operations on historical data of different types of objects, building models, providing market intelligent prediction for users from multiple dimensions, not only comprising traditional market prediction in a historical technical index mode, but also comprising market prediction based on information and transaction emotion analysis of the users, training to obtain a market prediction model based on technical indexes, a market prediction model based on information and a transaction emotion model based on user guessing, completing object market prediction by carrying out weight distribution on the three models, and obtaining a final prediction result by different weight weighted averages of different distribution weights of prediction accuracy rates of the three models, so that the prediction accuracy rate of the objects is effectively improved. The invention not only applies the object history technical index as the data acquisition object, but also brings the real-time information and the user emotion into the data acquisition category, thereby increasing the credibility and the accuracy of the model. According to the invention, the market situation intelligent prediction is provided for the user from multiple dimensions, more target object investment prediction information can be provided for the user, the user can know market situation analysis of the product from multiple directions and multiple dimensions, the user experience and the user viscosity are effectively improved, the user is further attracted, and the possibility is provided for realizing deep conversion of the user.
When the multi-model object market prediction method provided by the embodiment of the present invention is implemented specifically, in an embodiment, the method may include:
acquiring historical data of a target object;
preprocessing historical data of the object to determine preprocessed data;
extracting the characteristics of the preprocessed data, and establishing a prediction model;
dividing the preprocessed data into a training set and a testing set, and determining a market condition prediction model based on technical indexes, a market condition prediction model based on information and a transaction emotion model based on user guessing fluctuation by using the training set;
respectively determining the accuracy rates of the prediction results of a market forecasting model based on technical indexes, a market forecasting model based on information and a user guessing fluctuation transaction emotion model according to the test set;
according to the accuracy of the prediction results of the three models, weights are distributed to the market forecasting model based on the technical indexes, the market forecasting model based on the information and the trading emotion model based on the guessing of the fluctuation of the user, the market of the current object is predicted after weighted average, and the market forecasting result of the current object is determined.
The embodiment of the invention can provide more investment information for the user, so that the user can know the market trend analysis of the product in multiple directions and multiple dimensions, the user experience and the user viscosity are effectively improved, the user is further attracted, and the deep conversion of the user is possible. The invention uses the historical technical index data of the market objects in combination with the environmental factor data, and adopts a machine learning related algorithm to establish a model, and the model can provide intelligent forecast service of the market objects and provide related reference for investment decision of investors.
When the multi-model object market prediction method provided by the embodiment of the present invention is implemented specifically, in an embodiment, obtaining object history data includes:
and connecting a data acquisition system to obtain the historical data of the target object, wherein the historical data comprises the technical index of the target object, the basic data of the information and the data related to the guessing of the fluctuation of the user.
In the embodiment, as the market trend of the object has great uncertainty, the market trend is influenced not only by objective factors such as the historical trend and technical indexes of the object, but also by main factors such as public opinion and human emotion, so that the complexity of modeling and predicting the market trend by singly depending on the historical trend and the technical index data of the object and difficultly restoring the market trend prediction is realized, and the prediction accuracy is influenced; in order to solve the above problems, in the embodiment of the present invention, when acquiring and acquiring the historical data of the target object, the data acquisition system is connected to acquire the historical data of the target object including the technical index of the target object, the basic data of information, and the data related to the guessing of the fluctuation of the user; the invention not only applies the object history technical index as the data acquisition object, but also brings the real-time information and the user emotion into the data acquisition category, thereby increasing the credibility and the accuracy of the model.
When the multi-model object market prediction method provided by the embodiment of the invention is implemented specifically, in one embodiment, the obtaining of the object technical index includes:
connecting a data acquisition system, connecting a market information interface of the object to acquire technical indexes of the object; wherein, the subject matter technical index includes: the data acquisition system calculates and stores object cache data and technical index data of objects including spot goods, futures and stocks.
In the embodiment, when the technical index of the target object is acquired, the data acquisition system is connected, and the market information interface of the target object is connected to complete the acquisition of the technical index of the target object; wherein, the subject matter technical index includes: the object cache data and technical index data of the mark including spot goods, futures and stocks calculated and stored by the data acquisition system specifically comprise: the maximum price, the minimum price, the closing price, the opening price, the volume of bargain, the fluctuation range, MA5, MA10, MA30, KDJ, RSI, MACD, BOLL Bling band and other K lines and technical index data of 1 hour/4 hours/day/week/month.
When the multi-model object market prediction method provided by the embodiment of the invention is implemented, in one embodiment, the obtaining of the information basic data includes:
connecting a data acquisition system, and connecting a target object related information interface to obtain information basic data; wherein, the information basic data includes: historical financial reports, the influence of political events on precious metals and crude oil quotations, the discussion of quotations issued by users, transaction strategies issued by industry certification authorities, hot event analysis viewpoints and external quotation analysis information.
In the embodiment, when the information basic data is acquired, the data acquisition system is connected to the target object related information interface to complete the acquisition of the information basic data; wherein, the information basic data includes: historical financial reports, the influence of political events on precious metals and crude oil quotations, the discussion of quotations issued by users, transaction strategies issued by industry certification authorities, hot event analysis viewpoints and external quotation analysis information.
Because the market trend of the object has great uncertainty, it is not only influenced by objective factors such as the historical trend of the object itself, technical indicators, etc., but also influenced by subjective factors such as public opinion, human emotion, etc., therefore, the modeling prediction market is difficult to restore the complexity of the market forecast by solely depending on the historical trend of the object and the technical indicator data, and the accuracy of the forecast is influenced; the invention provides intelligent market forecast for the user from multiple dimensions, which not only comprises traditional market forecast through a historical technical index mode, but also comprises market forecast based on information and transaction emotion analysis of the user, and the market trend of the subject matter is greatly dependent on environmental factors, such as the influence of large environments of the emotion, politics, economy and the like of the trader. The invention not only applies the object history technical index as the data acquisition object, but also brings the real-time information and the user emotion into the data acquisition category, thereby increasing the credibility and the accuracy of the model.
In a specific implementation of the method for predicting the market quotation of the multi-model object according to the embodiment of the present invention, in an embodiment, the obtaining of the data related to the guessing of the fluctuation of the user includes:
and connecting a data acquisition system, butting a user historical transaction interface, and acquiring the user guessing rise and fall related data including the rise and fall results of the user.
In the embodiment, the trend of the user to the subject matter can guess the rise and fall, and in order to enhance the accuracy of model prediction, the data acquisition is carried out on the transaction emotion information of the user's ' rise ' or ' fall ', and the main process comprises the following steps: and connecting a data acquisition system, butting a user historical transaction interface, and acquiring the user guessing rise and fall related data including the rise and fall results of the user.
Because the market trend of the object has great uncertainty, it is not only influenced by objective factors such as the historical trend of the object itself, technical indicators, etc., but also influenced by subjective factors such as public opinion, human emotion, etc., therefore, the modeling prediction market is difficult to restore the complexity of the market forecast by solely depending on the historical trend of the object and the technical indicator data, and the accuracy of the forecast is influenced; the invention provides intelligent market forecast for the user from multiple dimensions, which not only comprises traditional market forecast through a historical technical index mode, but also comprises market forecast based on information and transaction emotion analysis of the user, and the market trend of the subject matter is greatly dependent on environmental factors, such as the influence of large environments of the emotion, politics, economy and the like of the trader. The invention not only applies the object history technical index as the data acquisition object, but also brings the real-time information and the user emotion into the data acquisition category, thereby increasing the credibility and the accuracy of the model.
When the multi-model object market prediction method provided by the embodiment of the present invention is implemented specifically, in an embodiment, the preprocessing is performed on object historical data to determine preprocessed data, and the method includes:
and preprocessing the technical indexes of the object, the basic data of the information and the data related to the guessing fluctuation of the user to obtain preprocessed data meeting the input standard of a machine learning algorithm.
In the embodiment, in order to enable the acquired historical data of the object to be used by machine learning, the technical index of the object, the basic data of information and the data related to the guessing and the fluctuating of the user need to be preprocessed to reach the input standard of a machine learning algorithm.
When the multi-model object market prediction method provided by the embodiment of the invention is implemented specifically, in one embodiment, feature extraction is performed on preprocessed data, and a prediction model is established, wherein the method comprises the following steps: and performing feature extraction on the preprocessed data by implementing feature engineering, and establishing a prediction model, wherein an algorithm for establishing the prediction model comprises a CNN (convolutional neural network) algorithm and a logistic regression algorithm in a machine learning algorithm. The forecasting model can provide intelligent forecasting service of the market quotation of the object and provide related reference for investment decision of investors.
Fig. 2 is a schematic process diagram of determining a market forecasting model based on technical indicators in a multi-model object market forecasting method according to an embodiment of the present invention, and as shown in fig. 2, when the multi-model object market forecasting method according to the embodiment of the present invention is implemented specifically, in an embodiment, determining a market forecasting model based on technical indicators includes:
step 201: according to the dimension of the time length to be predicted, inputting data by taking the characteristics of different dimensions;
step 202: according to feature input data of different dimensions, a CNN convolutional neural network algorithm is used for experimental parameter adjustment, a prediction model is trained by a training set, and a market condition prediction model based on technical indexes is determined through parameter adjustment learning.
In the embodiment, in order to train and obtain a market prediction model based on technical indexes, the main process comprises the following steps: firstly, according to the dimension of time length to be predicted, inputting data by taking the characteristics of different dimensions; and then inputting data according to features of different dimensions, performing experimental parameter adjustment by using a CNN convolutional neural network algorithm, training a prediction model by using a training set, and determining a market situation prediction model based on technical indexes through parameter adjustment learning.
In one example, characteristic input data of different dimensions are taken according to time dimensions (the rise and fall amplitude of the next day, the rise and fall amplitude of the next week and the rise and fall amplitude of the next month) to be predicted to perform a test, if the rise and fall amplitude of the next day is predicted, a day K line in a fixed time range is taken to perform analysis on corresponding technical indexes, the sampling frequency of historical data is subjected to test parameter adjustment, if the rise and fall amplitude of the next day is predicted, data of the previous 30 days can be sampled, and particularly, which points in the 30 days of sampling also need to be adjusted. The specific test algorithm adopts a CNN convolutional neural network, historical data are divided into a training set and a testing set, parameter adjustment learning is carried out, and finally a prediction model for the future market conditions of the target object based on technical indexes is generated.
When the multi-model object market prediction method provided by the embodiment of the invention is implemented, in one embodiment, the information basic data is preprocessed, and the method comprises the following steps:
labeling the information basic data by combining professional background knowledge, and determining preprocessed information basic data; wherein, the labeling process comprises: a piece of information is selected from the information basic data for labeling, and the selected piece of information can comprise a plurality of labels.
In an embodiment, in order to enable the information-based data to be used by the prediction model, the information-based data needs to be preprocessed, and the main processes include: analyzing the influence of historical financial and political events on precious metals and crude oil quotations, discussions on quotations issued by users, transaction strategies issued by network authentication large V, analysis views of hot events, external quotation analysis information, professional background knowledge and professional background knowledge, labeling information basic data and determining preprocessed information basic data; wherein, the labeling process comprises: a piece of information is selected from the information basic data for labeling, and the selected piece of information can comprise a plurality of labels.
Fig. 3 is a schematic process diagram of determining an information-based market forecasting model of a multi-model object market forecasting method according to an embodiment of the present invention, and as shown in fig. 3, when the multi-model object market forecasting method according to the embodiment of the present invention is implemented, in an embodiment, determining an information-based market forecasting model includes:
step 301: labeling the target object, and classifying the target object;
step 302: the method comprises the steps of using labels of information basic data as input data, using fluctuation of classes of object classification as learning targets, dividing preprocessed information basic data into a training set and a testing set, training a prediction model by combining a CNN convolutional neural network with the training set, and determining a market situation prediction model based on information.
In an embodiment, in order to train and obtain an information-based market prediction model, the main process includes: firstly, labeling the object, and classifying the object; and then, using the label of the information basic data as input data, using the fluctuation of the classification of the object as a learning target, dividing the preprocessed information basic data into a training set and a test set, training the prediction model by adopting a CNN convolutional neural network and combining the training set, and determining the market situation prediction model based on the information. In the embodiment, labeling processing is carried out on the target to be predicted, the target is classified, the label of the information is used as input, the fluctuation of the type of the target is used as a learning target, historical data is divided into a training set and a testing set, a CNN convolutional neural network is adopted for model training, and finally, a prediction model for the future market of the target based on the information is generated.
In one embodiment, the determining of the emotional model of the transaction based on the guessing of the fluctuation by the user includes:
extracting guess rise and fall transaction behavior results of each user participating in from guess rise and fall related data of the users;
setting initial weight by using the guess fluctuation transaction behavior result of each user as reference and the guess fluctuation accuracy and participation rate as the consideration data, modeling by adopting logistic regression algorithm,
and for the data with the accuracy and the participation rate higher than the preset threshold, dynamically adjusting the weight, regressing the history data of the user guessing the fluctuation, and determining a trading emotion model based on the user guessing the fluctuation.
In the embodiment, in order to train and obtain the emotion model based on the guessing of the fluctuation transaction of the user, the main process comprises the following steps: firstly, extracting guess rise and fall transaction behavior results of each user from guess rise and fall related data of the user; and then, taking the guess fluctuation transaction behavior result of each user as a reference, taking the guess fluctuation accuracy and the guess fluctuation participation rate as considered data, setting initial weight, modeling by adopting a logistic regression algorithm, and finally dynamically adjusting the weight for the data with the accuracy and the guess fluctuation participation rate higher than a preset threshold value, regressing the guess fluctuation history data of the user, and determining a sentiment model based on the guess fluctuation transaction of the user.
The trend of the user to the object can guess the fluctuation, the guess fluctuation result participated by each user is taken as the reference, the accuracy and the participation rate are taken as the considered data, the initial weight is set, the logistic regression algorithm is adopted for modeling, the weight is dynamically adjusted for the data with higher accuracy and participation rate, the reliability of the model is improved, and the historical data is regressed to improve the accuracy of the model.
In one embodiment, when the multi-model object market quotation prediction method provided by the embodiment of the invention is specifically implemented, the accuracy of the prediction result of each of the market quotation prediction model based on the technical indexes, the market quotation prediction model based on the information and the user guessing fluctuation transaction emotion model is respectively determined according to the test set, and the method comprises the following steps:
on the basis of the market forecasting model based on the technical indexes, the market forecasting model based on the information and the trading emotion model based on the user guessing whether to rise or fall obtained through training, the respective forecasting result accuracy rates of the three models are obtained through respectively inputting the test sets into the market forecasting model based on the technical indexes, the market forecasting model based on the information and the trading emotion model based on the user guessing whether to rise or fall.
In one embodiment, when the multi-model object market quotation prediction method provided by the embodiment of the present invention is specifically implemented, according to the accuracy of the prediction results of the three models, weights are assigned to the market quotation prediction model based on the technical indexes, the market quotation prediction model based on the information, and the user guessing fluctuation transaction emotion model, the current object market quotation is predicted after weighted averaging, and the current object market quotation prediction result is determined, including:
according to the accuracy rates of the prediction results of the market forecasting model based on the technical indexes, the market forecasting model based on the information and the transaction emotion model based on the user guessing fluctuation, different weights are distributed to the market forecasting model based on the technical indexes, the market forecasting model based on the information and the transaction emotion model based on the user guessing fluctuation, the higher the accuracy rate is, the larger the weight is, the comprehensive forecasting model for finally forecasting the current object market is obtained through weighted evaluation, the current object market is forecasted through the comprehensive forecasting model, and the current object market forecasting result is determined.
And distributing different weights to the final prediction result according to the accuracy of the three models, and performing weighted average to obtain the final prediction result.
The invention provides intelligent market prediction for the user from multiple dimensions, and not only comprises traditional market prediction through a historical technical index mode, but also comprises market prediction based on information and transaction emotion analysis of the user. The traditional market forecasting system is usually only dependent on the historical technical index data of the subject matter, but neglects the market trend which is greatly dependent on the environmental factors, such as the influence of the large environment of the emotion, politics, economy and the like of the trader. The invention not only applies the object history technical index as the data acquisition object, but also brings the real-time information and the user emotion into the data acquisition category, thereby increasing the credibility and the accuracy of the model.
The method can provide more investment information for the user, so that the user can know market trend analysis of the product in multiple directions and multiple dimensions, user experience and user stickiness are effectively improved, the user is further attracted, and possibility is provided for realizing deep conversion of the user. The invention uses the historical technical index data of the market objects in combination with the environmental factor data, and adopts a machine learning related algorithm to establish a model, and the model can provide intelligent forecast service of the market objects and provide related reference for investment decision of investors.
In the process of predicting the object, the method not only considers the historical technical indexes of the object, but also considers the near term public opinion information, trader emotion information and other multidimensional information, carries out data preprocessing, characteristic engineering and other operations on different types of information, finally builds a model to complete prediction, obtains the final prediction result by different weight weighted averages of different distribution of prediction accuracy rates of 3 models, and effectively improves the prediction accuracy rate of the object.
The multi-model object market prediction method provided by the embodiment of the invention is briefly described below with reference to specific scenarios:
the multi-model object market quotation prediction method provided by the embodiment of the invention can be used for building an intelligent transaction market quotation prediction system, providing more investment information for users, enabling the users to know market quotation trend analysis of products in a multi-azimuth and multi-dimension manner, effectively improving user experience and user stickiness, further attracting the users and providing possibility for realizing deep conversion of the users. The invention uses the historical technical index data of the market objects in combination with the environmental factor data, and adopts a machine learning related algorithm to establish a model, and the model can provide intelligent forecast service of the market objects and provide related reference for investment decision of investors.
The method mainly comprises three steps of data acquisition and preprocessing, feature extraction and model training and prediction result providing.
Wherein the data acquisition and preprocessing mainly comprises:
1. completing historical technical index collection by a market information interface of the object;
2. the information related information interface of the target is connected to complete the information related information acquisition;
3. data acquisition is carried out on the transaction emotion information of the 'look up' or 'look down' of the user;
4. carrying out data preprocessing on the acquired data so as to carry out modeling by adopting a relevant machine learning algorithm;
the characteristic extraction and model training part mainly comprises:
1. processing and extracting characteristic input required by the following 3 models;
2. training a market forecasting model based on technical indexes such as KDJ and MA;
3. training a market prediction model based on financial information;
4. predicting market quotations based on a user's guess fluctuation' trading emotion building model;
and dividing historical data into a training set and a testing set, and after model training is completed, distributing different weights for the 3 dimensions according to the prediction accuracy of the 3 dimensions in the testing set.
Wherein, providing the prediction result mainly comprises:
the final prediction result is obtained by performing weighted average on the prediction values of the 3 dimensions.
The invention provides intelligent market prediction for the user from multiple dimensions, and not only comprises traditional market prediction through a historical technical index mode, but also comprises market prediction based on information and transaction emotion analysis of the user. The traditional market forecasting system is usually only dependent on the historical technical index data of the subject matter, but neglects the market trend which is greatly dependent on the environmental factors, such as the influence of the large environment of the emotion, politics, economy and the like of the trader. The invention not only applies the object history technical index as the data acquisition object, but also brings the real-time information and the user emotion into the data acquisition category, thereby increasing the credibility and the accuracy of the model.
Firstly, technical indexes of a target object, basic data of information and data related to guessing and fluctuating of a user need to be acquired by technicians. Second, data preprocessing is required on these data to meet the requirements that can be input as a machine learning related algorithm. Thirdly, a proper machine learning algorithm is needed to be selected for modeling and parameter adjustment. And finally, distributing corresponding weight to the three data according to the test accuracy rates with different data dimensions, and carrying out weighted average to obtain a final prediction result.
The most important steps of the method are a data preprocessing stage and a modeling stage, and the two stages directly determine the accuracy of model prediction.
The embodiment of the invention also provides a modular example of the multi-model object market prediction method, which mainly comprises the following steps:
the system comprises three modules of data acquisition and preprocessing, feature extraction and model training and prediction result output, wherein the data acquisition and preprocessing, feature extraction and model training modules have different data processing modes aiming at three dimensions, the prediction result output module distributes weights to the models with different dimensions according to the accuracy of the prediction results of the models with different dimensions, and the final prediction result is obtained by weighted average. The data acquisition, preprocessing, feature extraction and model training methods of different dimensions of the subject matter are explained in detail.
1. Intelligent analysis of market trend based on technical indexes such as KDJ, MA and the like:
the historical data of the labels such as spot goods, futures, stocks and the like calculated and stored by the data acquisition system specifically comprise K lines and technical index data of 1 hour/4 hour/day/week/month, the highest price, the lowest price, the closing price, the opening price, the volume of transaction, the fluctuation range, MA5, MA10, MA30, KDJ, RSI, MACD, BOLL Bolin belt and the like.
And (3) inputting characteristic input data of different dimensions according to the time dimension (the rise and fall amplitude of the next day, the rise and fall amplitude of the next week and the rise and fall amplitude of the next month) to be predicted, analyzing the corresponding technical index of the day K line in a fixed time range if the rise and fall amplitude of the next day is predicted, performing test reference adjustment on the sampling frequency of historical data, sampling the data of the previous 30 days if the rise and fall amplitude of the next day is predicted, and specifically adjusting which points in the 30 days of sampling.
The specific test algorithm adopts a CNN convolutional neural network, historical data are divided into a training set and a testing set, parameter adjusting learning is carried out, and finally a prediction model of the future market conditions of the target object is generated.
2. The market trend intelligent analysis of the information basic data accessed through the information source is as follows:
the method comprises the steps of analyzing the influence of historical finance and political events on precious metals and crude oil quotations, discussing quotations issued by users, analyzing trading strategies issued by large V and hotspot events, analyzing information of external quotations, combining professional background knowledge, preprocessing collected data, wherein the preprocessing specifically comprises labeling certain section of information, the section of information can comprise a plurality of labels, labeling the objects to be predicted, classifying the objects, using the labels of the information as input, using the fluctuation of the class of the objects as a learning target, dividing historical data into a training set and a testing set, performing model training by adopting a CNN convolutional neural network, and finally generating a prediction model of the future quotations of the objects.
3. Intelligent prediction of market based on user transaction emotion:
the trend of the user to the object can guess the fluctuation, the guess fluctuation result participated by each user is taken as the reference, the accuracy and the participation rate are taken as the considered data, the initial weight is set, the logistic regression algorithm is adopted for modeling, the weight is dynamically adjusted for the data with higher accuracy and participation rate, the reliability of the model is improved, and the historical data is regressed to improve the accuracy of the model.
And distributing different weights to the final prediction result according to the accuracy of the three models, and performing weighted average to obtain the final prediction result.
The modular example of the multi-model object market prediction method mainly comprises a data processing process, wherein the data processing process comprises the following steps:
1. the data acquisition module acquires data required by the model 1, the model 2 and the model 3;
2. the data preprocessing module preprocesses the data acquired in the step 1 and mainly comprises data standardization processing and the like;
3. performing feature extraction and processing on the data standardized in the step 2, wherein the feature extraction and processing mainly comprises data processing such as principal component analysis;
4. dividing historical data into a training set test set, and respectively training to obtain a model 1, a model 2 and a model 3;
5. according to the prediction accuracy of the model 1, the model 2 and the model 3 in the test set, different weights are distributed to the models, the higher the accuracy is, the larger the weight is, and finally the final prediction result is obtained through weighted average.
The core innovation points of the embodiment of the invention comprise:
1. the subject matter market prediction data acquisition range;
2. applying different characteristic engineering and modeling methods aiming at different data types;
3. environmental factors influencing the market quotation of the object, such as information, the emotion of a trader and the like, are quantized and input into the model for learning, so that the accuracy of model prediction is improved;
4. the method comprises the steps of quantifying environmental factor data such as target object technical index data, information, trader emotion and the like, respectively establishing prediction models, finally distributing different weights according to the prediction accuracy of each model, obtaining a final result through weighted average, and improving the prediction accuracy of the final model.
In the process of predicting the object, the method not only considers the historical technical indexes of the object, but also considers the near term public opinion information, trader emotion information and other multidimensional information, carries out data preprocessing, characteristic engineering and other operations on different types of information, finally builds a model to complete prediction, obtains the final prediction result by different weight weighted averages of different distribution of prediction accuracy rates of 3 models, and effectively improves the prediction accuracy rate of the object.
Fig. 4 is a schematic diagram of a computer device for executing a multi-model object market forecasting method implemented by the present invention, and as shown in fig. 4, an embodiment of the present invention further provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the multi-model object market forecasting method when executing the computer program.
An embodiment of the present invention further provides a computer-readable storage medium, where a computer program for implementing the multi-model object market prediction method is stored in the computer-readable storage medium.
The embodiment of the invention also provides a multi-model object market prediction device, which is described in the following embodiment. Because the principle of the device for solving the problems is similar to the multi-model object market forecasting method, the implementation of the device can refer to the implementation of the multi-model object market forecasting method, and repeated parts are not repeated.
Fig. 5 is a schematic diagram of a multi-model object market forecasting apparatus according to an embodiment of the present invention, and as shown in fig. 5, an embodiment of the present invention further provides a multi-model object market forecasting apparatus, which may include:
a target historical data acquiring module 501, configured to acquire target historical data;
a preprocessing module 502, configured to preprocess the historical data of the object and determine preprocessed data;
a feature extraction module 503, configured to perform feature extraction on the preprocessed data, and establish a prediction model;
the model training module 504 is used for dividing the preprocessed data into a training set and a test set, and determining a market forecasting model based on technical indexes, a market forecasting model based on information and a transaction emotion model based on user guessing fluctuation by using the training set;
a prediction result accuracy determining module 505, configured to determine, according to the test set, respective prediction result accuracy of a market forecasting model based on the technical indicators, a market forecasting model based on the information, and a user guessing fluctuation transaction emotion model;
and the current target object market prediction module 506 is used for distributing weights to the market prediction model based on the technical indexes, the market prediction model based on the information and the user guessing fluctuation transaction emotion model according to the accuracy of the prediction results of the three models, predicting the current target object market after weighted averaging, and determining the current target object market prediction result.
In an embodiment of the invention, when the multi-model object market prediction apparatus provided in the embodiment of the present invention is implemented specifically, the object history data obtaining module is specifically configured to:
and connecting a data acquisition system to obtain the historical data of the target object, wherein the historical data comprises the technical index of the target object, the basic data of the information and the data related to the guessing of the fluctuation of the user.
In an embodiment of the invention, when the multi-model object market prediction apparatus provided in the embodiment of the present invention is implemented specifically, the object history data obtaining module is further configured to:
connecting a data acquisition system, connecting a market information interface of the object to acquire technical indexes of the object; wherein, the subject matter technical index includes: the data acquisition system calculates and stores object cache data and technical index data of objects including spot goods, futures and stocks.
In an embodiment of the invention, when the multi-model object market prediction apparatus provided in the embodiment of the present invention is implemented specifically, the object history data obtaining module is further configured to:
connecting a data acquisition system, and connecting a target object related information interface to obtain information basic data; wherein, the information basic data includes: historical financial reports, the influence of political events on precious metals and crude oil quotations, the discussion of quotations issued by users, transaction strategies issued by industry certification authorities, hot event analysis viewpoints and external quotation analysis information.
In an embodiment of the invention, when the multi-model object market prediction apparatus provided in the embodiment of the present invention is implemented specifically, the object history data obtaining module is further configured to:
and connecting a data acquisition system, butting a user historical transaction interface, and acquiring the user guessing rise and fall related data including the rise and fall results of the user.
In an embodiment of the invention, when the multi-model object market prediction apparatus provided in the embodiment of the present invention is implemented specifically, the preprocessing module is specifically configured to:
and preprocessing the technical indexes of the object, the basic data of the information and the data related to the guessing fluctuation of the user to obtain preprocessed data meeting the input standard of a machine learning algorithm.
In an embodiment of the invention, when the multi-model object market prediction apparatus provided in the embodiment of the present invention is implemented specifically, the model training module is specifically configured to:
according to the dimension of the time length to be predicted, inputting data by taking the characteristics of different dimensions;
according to feature input data of different dimensions, a CNN convolutional neural network algorithm is used for experimental parameter adjustment, a prediction model is trained by a training set, and a market condition prediction model based on technical indexes is determined through parameter adjustment learning.
In an embodiment of the invention, when the multi-model object market prediction apparatus provided in the embodiment of the present invention is implemented specifically, the preprocessing module is further configured to:
labeling the information basic data by combining professional background knowledge, and determining preprocessed information basic data; wherein, the labeling process comprises: a piece of information is selected from the information basic data for labeling, and the selected piece of information can comprise a plurality of labels.
In an embodiment of the invention, when the multi-model object market prediction device provided in the embodiment of the present invention is implemented specifically, the model training module is further configured to:
labeling the target object, and classifying the target object;
the method comprises the steps of using labels of information basic data as input data, using fluctuation of classes of object classification as learning targets, dividing preprocessed information basic data into a training set and a testing set, training a prediction model by combining a CNN convolutional neural network with the training set, and determining a market situation prediction model based on information.
In an embodiment of the invention, when the multi-model object market prediction device provided in the embodiment of the present invention is implemented specifically, the model training module is further configured to:
extracting guess rise and fall transaction behavior results of each user participating in from guess rise and fall related data of the users;
setting initial weight by using the guess fluctuation transaction behavior result of each user as reference and the guess fluctuation accuracy and participation rate as the consideration data, modeling by adopting logistic regression algorithm,
and for the data with the accuracy and the participation rate higher than the preset threshold, dynamically adjusting the weight, regressing the history data of the user guessing the fluctuation, and determining a trading emotion model based on the user guessing the fluctuation.
To sum up, the method and the device for predicting the market quotation of the multi-model object provided by the embodiment of the invention comprise the following steps: firstly, acquiring historical data of a target object; then, preprocessing the historical data of the object to determine preprocessed data; then, extracting the characteristics of the preprocessed data, and establishing a prediction model; continuously dividing the preprocessed data into a training set and a testing set, and determining a market condition prediction model based on technical indexes, a market condition prediction model based on information and a transaction emotion model based on user guessing fluctuation by using the training set; next, respectively determining the accuracy of the prediction results of a market forecasting model based on technical indexes, a market forecasting model based on information and a user guessing fluctuation transaction emotion model according to the test set; and finally, according to the accuracy of the prediction results of the three models, distributing weights to the market quotation prediction model based on the technical indexes, the market quotation prediction model based on the information and the user guessing fluctuation transaction emotion model, predicting the market quotation of the current target after weighted averaging, and determining the market quotation prediction result of the current target. The method comprises the steps of carrying out data preprocessing, characteristic engineering and other operations on historical data of different types of objects, building models, providing market intelligent prediction for users from multiple dimensions, not only comprising traditional market prediction in a historical technical index mode, but also comprising market prediction based on information and transaction emotion analysis of the users, training to obtain a market prediction model based on technical indexes, a market prediction model based on information and a transaction emotion model based on user guessing, completing object market prediction by carrying out weight distribution on the three models, and obtaining a final prediction result by different weight weighted averages of different distribution weights of prediction accuracy rates of the three models, so that the prediction accuracy rate of the objects is effectively improved. The invention not only applies the object history technical index as the data acquisition object, but also brings the real-time information and the user emotion into the data acquisition category, thereby increasing the credibility and the accuracy of the model. According to the invention, the market situation intelligent prediction is provided for the user from multiple dimensions, more target object investment prediction information can be provided for the user, the user can know market situation analysis of the product from multiple directions and multiple dimensions, the user experience and the user viscosity are effectively improved, the user is further attracted, and the possibility is provided for realizing deep conversion of the user.
The invention provides intelligent market prediction for the user from multiple dimensions, and not only comprises traditional market prediction through a historical technical index mode, but also comprises market prediction based on information and transaction emotion analysis of the user. The traditional market forecasting system is usually only dependent on the historical technical index data of the subject matter, but neglects the market trend which is greatly dependent on the environmental factors, such as the influence of the large environment of the emotion, politics, economy and the like of the trader. The invention not only applies the object history technical index as the data acquisition object, but also brings the real-time information and the user emotion into the data acquisition category, thereby increasing the credibility and the accuracy of the model. The method can provide more investment information for the user, so that the user can know market trend analysis of the product in multiple directions and multiple dimensions, user experience and user stickiness are effectively improved, the user is further attracted, and possibility is provided for realizing deep conversion of the user. The invention uses the historical technical index data of the market objects in combination with the environmental factor data, and adopts a machine learning related algorithm to establish a model, and the model can provide intelligent forecast service of the market objects and provide related reference for investment decision of investors.
In the process of predicting the object, the method not only considers the historical technical indexes of the object, but also considers the near term public opinion information, trader emotion information and other multidimensional information, carries out data preprocessing, characteristic engineering and other operations on different types of information, finally builds a model to complete prediction, obtains the final prediction result by different weight weighted averages of different distribution of prediction accuracy rates of 3 models, and effectively improves the prediction accuracy rate of the object.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (22)

1. A multi-model object market prediction method is characterized by comprising the following steps:
acquiring historical data of a target object;
preprocessing historical data of the object to determine preprocessed data;
extracting the characteristics of the preprocessed data, and establishing a prediction model;
dividing the preprocessed data into a training set and a testing set, and determining a market condition prediction model based on technical indexes, a market condition prediction model based on information and a transaction emotion model based on user guessing fluctuation by using the training set;
respectively determining the accuracy rates of the prediction results of a market forecasting model based on technical indexes, a market forecasting model based on information and a user guessing fluctuation transaction emotion model according to the test set;
according to the accuracy of the prediction results of the three models, weights are distributed to the market forecasting model based on the technical indexes, the market forecasting model based on the information and the trading emotion model based on the guessing of the fluctuation of the user, the market of the current object is predicted after weighted average, and the market forecasting result of the current object is determined.
2. The method of claim 1, wherein obtaining subject matter history data comprises:
and connecting a data acquisition system to obtain the historical data of the target object, wherein the historical data comprises the technical index of the target object, the basic data of the information and the data related to the guessing of the fluctuation of the user.
3. The method of claim 2, wherein obtaining a subject technical indicator comprises:
connecting a data acquisition system, connecting a market information interface of the object to acquire technical indexes of the object; wherein, the subject matter technical index includes: the data acquisition system calculates and stores object cache data and technical index data of objects including spot goods, futures and stocks.
4. The method of claim 2, wherein obtaining information-based data comprises:
connecting a data acquisition system, and connecting a target object related information interface to obtain information basic data; wherein, the information basic data includes: historical financial reports, the influence of political events on precious metals and crude oil quotations, the discussion of quotations issued by users, transaction strategies issued by industry certification authorities, hot event analysis viewpoints and external quotation analysis information.
5. The method of claim 2, wherein obtaining data related to the user's guessing at the rise or fall comprises:
and connecting a data acquisition system, butting a user historical transaction interface, and acquiring the user guessing rise and fall related data including the rise and fall results of the user.
6. The method of claim 2, wherein pre-processing the subject matter history data to determine pre-processed data comprises:
and preprocessing the technical indexes of the object, the basic data of the information and the data related to the guessing fluctuation of the user to obtain preprocessed data meeting the input standard of a machine learning algorithm.
7. The method of claim 6, wherein determining a technology indicator based market prediction model comprises:
according to the dimension of the time length to be predicted, inputting data by taking the characteristics of different dimensions;
according to feature input data of different dimensions, a CNN convolutional neural network algorithm is used for experimental parameter adjustment, a prediction model is trained by a training set, and a market condition prediction model based on technical indexes is determined through parameter adjustment learning.
8. The method of claim 6, wherein preprocessing the information-based data comprises:
labeling the information basic data by combining professional background knowledge, and determining preprocessed information basic data; wherein, the labeling process comprises: a piece of information is selected from the information basic data for labeling, and the selected piece of information can comprise a plurality of labels.
9. The method of claim 8, wherein determining an information-based market prediction model comprises:
labeling the target object, and classifying the target object;
the method comprises the steps of using labels of information basic data as input data, using fluctuation of classes of object classification as learning targets, dividing preprocessed information basic data into a training set and a testing set, training a prediction model by combining a CNN convolutional neural network with the training set, and determining a market situation prediction model based on information.
10. The method of claim 6, wherein determining the emotion model based on the user guessing the transaction comprises:
extracting guess rise and fall transaction behavior results of each user participating in from guess rise and fall related data of the users;
setting initial weight by using the guess fluctuation transaction behavior result of each user as reference and the guess fluctuation accuracy and participation rate as the consideration data, modeling by adopting logistic regression algorithm,
and for the data with the accuracy and the participation rate higher than the preset threshold, dynamically adjusting the weight, regressing the history data of the user guessing the fluctuation, and determining a trading emotion model based on the user guessing the fluctuation.
11. A multi-model object market prediction device, comprising:
the system comprises a target historical data acquisition module, a data processing module and a data processing module, wherein the target historical data acquisition module is used for acquiring target historical data;
the preprocessing module is used for preprocessing the historical data of the object and determining preprocessed data;
the characteristic extraction module is used for extracting the characteristics of the preprocessed data and establishing a prediction model;
the model training module is used for dividing the preprocessed data into a training set and a testing set, and determining a market forecasting model based on technical indexes, a market forecasting model based on information and a user guessing fluctuation transaction emotion model by using the training set;
the prediction result accuracy rate determining module is used for respectively determining the respective prediction result accuracy rates of the market condition prediction model based on the technical indexes, the market condition prediction model based on the information and the user guessing fluctuation transaction emotion model according to the test set;
and the current target object market forecasting module is used for distributing weights to the market forecasting model based on the technical indexes, the market forecasting model based on the information and the user guessing fluctuation transaction emotion model according to the respective forecasting result accuracy of the three models, forecasting the current target object market after weighted averaging, and determining the current target object market forecasting result.
12. The apparatus of claim 11, wherein the subject matter history data acquisition module is specifically configured to:
and connecting a data acquisition system to obtain the historical data of the target object, wherein the historical data comprises the technical index of the target object, the basic data of the information and the data related to the guessing of the fluctuation of the user.
13. The apparatus of claim 12, wherein the subject matter history data acquisition module is further configured to:
connecting a data acquisition system, connecting a market information interface of the object to acquire technical indexes of the object; wherein, the subject matter technical index includes: the data acquisition system calculates and stores object cache data and technical index data of objects including spot goods, futures and stocks.
14. The apparatus of claim 12, wherein the subject matter history data acquisition module is further configured to:
connecting a data acquisition system, and connecting a target object related information interface to obtain information basic data; wherein, the information basic data includes: historical financial reports, the influence of political events on precious metals and crude oil quotations, the discussion of quotations issued by users, transaction strategies issued by industry certification authorities, hot event analysis viewpoints and external quotation analysis information.
15. The apparatus of claim 12, wherein the subject matter history data acquisition module is further configured to:
and connecting a data acquisition system, butting a user historical transaction interface, and acquiring the user guessing rise and fall related data including the rise and fall results of the user.
16. The apparatus of claim 12, wherein the pre-processing module is specifically configured to:
and preprocessing the technical indexes of the object, the basic data of the information and the data related to the guessing fluctuation of the user to obtain preprocessed data meeting the input standard of a machine learning algorithm.
17. The apparatus of claim 16, wherein the model training module is specifically configured to:
according to the dimension of the time length to be predicted, inputting data by taking the characteristics of different dimensions;
according to feature input data of different dimensions, a CNN convolutional neural network algorithm is used for experimental parameter adjustment, a prediction model is trained by a training set, and a market condition prediction model based on technical indexes is determined through parameter adjustment learning.
18. The apparatus of claim 16, wherein the pre-processing module is further configured to:
labeling the information basic data by combining professional background knowledge, and determining preprocessed information basic data; wherein, the labeling process comprises: a piece of information is selected from the information basic data for labeling, and the selected piece of information can comprise a plurality of labels.
19. The apparatus of claim 18, wherein the model training module is further configured to:
labeling the target object, and classifying the target object;
the method comprises the steps of using labels of information basic data as input data, using fluctuation of classes of object classification as learning targets, dividing preprocessed information basic data into a training set and a testing set, training a prediction model by combining a CNN convolutional neural network with the training set, and determining a market situation prediction model based on information.
20. The apparatus of claim 16, wherein the model training module is further configured to:
extracting guess rise and fall transaction behavior results of each user participating in from guess rise and fall related data of the users;
setting initial weight by using the guess fluctuation transaction behavior result of each user as reference and the guess fluctuation accuracy and participation rate as the consideration data, modeling by adopting logistic regression algorithm,
and for the data with the accuracy and the participation rate higher than the preset threshold, dynamically adjusting the weight, regressing the history data of the user guessing the fluctuation, and determining a trading emotion model based on the user guessing the fluctuation.
21. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method of any of claims 1 to 10 when executing the computer program.
22. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program for executing a method according to any one of claims 1 to 10.
CN202110780081.1A 2021-07-09 2021-07-09 Multi-model object market quotation prediction method and device Pending CN113449923A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110780081.1A CN113449923A (en) 2021-07-09 2021-07-09 Multi-model object market quotation prediction method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110780081.1A CN113449923A (en) 2021-07-09 2021-07-09 Multi-model object market quotation prediction method and device

Publications (1)

Publication Number Publication Date
CN113449923A true CN113449923A (en) 2021-09-28

Family

ID=77815654

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110780081.1A Pending CN113449923A (en) 2021-07-09 2021-07-09 Multi-model object market quotation prediction method and device

Country Status (1)

Country Link
CN (1) CN113449923A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116228424A (en) * 2023-03-03 2023-06-06 深圳市深弈科技有限公司 Simulation matching method based on migration learning optimization

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116228424A (en) * 2023-03-03 2023-06-06 深圳市深弈科技有限公司 Simulation matching method based on migration learning optimization
CN116228424B (en) * 2023-03-03 2024-06-04 深圳市深弈科技有限公司 Simulation matching method based on migration learning optimization

Similar Documents

Publication Publication Date Title
Tsantekidis et al. Forecasting stock prices from the limit order book using convolutional neural networks
CN110070391B (en) Data processing method and device, computer readable medium and electronic equipment
CN107766929B (en) Model analysis method and device
CN110400022B (en) Cash consumption prediction method and device for self-service teller machine
CN106909933B (en) A kind of stealing classification Forecasting Methodology of three stages various visual angles Fusion Features
CN110222880B (en) Service risk determining method, model training method and data processing method
CN109583966B (en) High-value customer identification method, system, equipment and storage medium
CN110147389A (en) Account number treating method and apparatus, storage medium and electronic device
CN109636482B (en) Data processing method and system based on similarity model
CN113469730A (en) Customer repurchase prediction method and device based on RF-LightGBM fusion model under non-contract scene
CN111192133A (en) Method and device for generating risk model after user loan and electronic equipment
CN111199469A (en) User payment model generation method and device and electronic equipment
CN114202243A (en) Engineering project management risk early warning method and system based on random forest
CN112132679A (en) Product recommendation method and device
CN113421168B (en) Intelligent machining system for mechanical basic parts
Wimmer et al. Leveraging vision-language models for granular market change prediction
CN113449923A (en) Multi-model object market quotation prediction method and device
CN113420909A (en) User response information prediction model establishing method and information prediction method
CN111695625A (en) Mobile application popularity prediction method based on multi-level attention mechanism network
CN110738565A (en) Real estate finance artificial intelligence composite wind control model based on data set
CN114626940A (en) Data analysis method and device and electronic equipment
CN113987351A (en) Artificial intelligence based intelligent recommendation method and device, electronic equipment and medium
CN110264251B (en) Data organization form for representing cash flow and prediction method based on multi-task learning
CN113850609A (en) Customer management system, method, computer equipment and storage medium
JP2023516035A (en) A method and system for processing data with varying temporal characteristics to generate predictions about management arrangements using a random forest classifier

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination