WO2023185125A1 - Procédé et appareil de traitement de données de ressource produit, dispositif électronique et support de stockage - Google Patents

Procédé et appareil de traitement de données de ressource produit, dispositif électronique et support de stockage Download PDF

Info

Publication number
WO2023185125A1
WO2023185125A1 PCT/CN2022/140756 CN2022140756W WO2023185125A1 WO 2023185125 A1 WO2023185125 A1 WO 2023185125A1 CN 2022140756 W CN2022140756 W CN 2022140756W WO 2023185125 A1 WO2023185125 A1 WO 2023185125A1
Authority
WO
WIPO (PCT)
Prior art keywords
target
resources
features
product
product resources
Prior art date
Application number
PCT/CN2022/140756
Other languages
English (en)
Chinese (zh)
Inventor
刘瀚文
Original Assignee
富途网络科技(深圳)有限公司
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 富途网络科技(深圳)有限公司 filed Critical 富途网络科技(深圳)有限公司
Publication of WO2023185125A1 publication Critical patent/WO2023185125A1/fr

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
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • 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"
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Definitions

  • This application relates to the field of artificial intelligence technology, specifically, to a data processing method and device for product resources, electronic equipment, and computer-readable storage media.
  • Product resources are usually displayed in corresponding applications.
  • corresponding pages are set up to display some data related to product resources. Users can use these Relevant information is obtained from the data. More professional users can extract their personal understanding of the product resources from the displayed data, thereby making corresponding decisions about the product resources.
  • professional users have higher requirements for their personal abilities. ; However, for most non-professional users, they cannot effectively use this data to make relatively correct decisions at a favorable time point, resulting in low information volume and information transmission efficiency of product resources.
  • embodiments of the present application provide a data processing method and device for product resources, electronic equipment, and computer-readable storage media, aiming to solve the problem of the amount of information on product resources when making decisions about product resources. and low efficiency of information transmission.
  • a data processing method for product resources including:
  • the target interpretation information of the target product resources is obtained.
  • a data processing device for product resources including:
  • the extraction module is configured to extract original features from the original data of the target product resources, and perform feature construction processing based on the original features to obtain the target features;
  • the first determination module is configured to determine the predicted revenue probability value of the target product resource based on the target characteristics
  • the second determination module is configured to determine the trend of the target product resource based on the predicted profit probability value of the target product resource and the predicted profit probability value of other product resources;
  • the target interpretation information module is configured to obtain the target interpretation information of the target product resources based on the trend and the original interpretation information of the target product resources.
  • an electronic device including: one or more processors; a storage device for storing one or more programs. When one or more programs are processed by one or more processors, When executed, the electronic device is caused to implement the previous data processing method of product resources.
  • a computer-readable storage medium on which computer-readable instructions are stored.
  • the computer-readable instructions are executed by a processor of a computer, the computer is caused to execute the above product resource data.
  • Approach the computer-readable instructions
  • a computer program product or a computer program is provided, the computer program product or the computer program including computer instructions, and the computer instructions are stored in a computer-readable storage medium.
  • the processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the data processing method of product resources provided in the various optional embodiments described above.
  • the original features of the target product resources are constructed to obtain the target features.
  • the target features have more information than the original features.
  • the target product resources are determined based on the target features.
  • the predicted revenue probability value can be more accurate; at the same time, the trend is also obtained based on the preset revenue probability value, and the target interpretation information is obtained based on the trend trend and the original interpretation information.
  • the user understands the future direction of the target product resources by viewing the target interpretation information.
  • Figure 1 is a schematic diagram of an implementation environment involved in this application.
  • Figure 2 is a flow chart of a data processing method for product resources in one embodiment of this application.
  • FIG. 3 is a flow chart of step S210 in one embodiment of the present application.
  • Figure 4 is a flow chart of a data processing method for product resources in another embodiment of the present application.
  • FIG. 5 is a flow chart also included after step S440 in an embodiment of the present application.
  • FIG. 6 is a flow chart also included before step S420 in one embodiment of the present application.
  • FIG. 7 is a flow chart of step S620 in one embodiment of the present application.
  • Figure 8 is a flow chart of step S240 in one embodiment of the present application.
  • Figure 9 is a flow chart of a data processing method for product resources in another embodiment of the present application.
  • Figure 10 is a block diagram of a data processing device for product resources involved in this application.
  • FIG. 11 is a schematic structural diagram of a computer system suitable for implementing an electronic device according to an embodiment of the present application.
  • FIG. 1 is a schematic diagram of an implementation environment involved in this application.
  • the implementation environment includes a terminal 110 and a server 120.
  • the terminal 110 and the server 120 communicate through a wired or wireless network.
  • the terminal 110 runs an application program related to product resources, and the user can view related product resources or perform corresponding operations on the product resources on the application program.
  • the application is provided with corresponding pages for displaying various data of product resources, and at the same time, corresponding interpretation pages are provided for displaying the target interpretation information obtained after the product resources are processed by the data processing method.
  • the terminal 110 may be a smartphone, a tablet, a laptop, a computer, or any other electronic device that can run applications related to product resources.
  • the server 120 stores a large amount of product resource related data, and the terminal obtains the related data from the server when it needs it.
  • Server 120 may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or may provide cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, Cloud servers for basic cloud computing services such as middleware services, domain name services, security services, CDN (Content Delivery Network, content distribution network), and big data and artificial intelligence platforms are not restricted here.
  • product resources include but are not limited to virtual product resources, such as stocks, futures, options, securities, virtual currency, funds, or foreign exchange, etc.
  • the data processing method for product resources provided by the embodiment of this application is specifically executed by the terminal 110 in the embodiment environment shown in Figure 1.
  • the data processing method of product resources provided by this application constructs features of the original characteristics of the target product resource to obtain the target Features, target features have more information than original features, and it can be more accurate to determine the predicted profit probability value of target product resources based on target features; at the same time, the trend is also obtained based on the preset profit probability value, based on the trend and
  • the target interpretation information is obtained from the original interpretation information.
  • the user understands the future direction of the target product resources by viewing the target interpretation information. When making corresponding decisions about the target product resources, the target interpretation information can be combined with the target interpretation information without relying on human experience.
  • FIG. 2 is a flow chart of a data processing method for product resources according to an exemplary embodiment. The method can be applied to the implementation environment shown in FIG. 1 . As shown in Figure 2, in an exemplary embodiment, the data processing method of the product resource may include steps S210 to S240, which are described in detail as follows:
  • Step S210 Extract original features from the original data of the target product resources, and perform feature construction processing based on the original features to obtain the target features.
  • the details page displays data of the target product resources in different dimensions. These data in different dimensions are the original data of the target product resources. Data, the original data includes a large amount of data, but some data is meaningless. Therefore, it is necessary to extract original features from the original data that can be used to predict future profits.
  • Feature construction refers to the process of constructing new features from original features to generate new features that can better reflect business characteristics. These new features must be closely related to the desired predicted revenue probability value.
  • step S210 feature construction processing is performed based on the original features to obtain the target features, including steps S310 to S330.
  • the details are as follows:
  • Step S310 Perform feature intersection processing on multiple original features to obtain intersection features.
  • feature cross processing is performed on each original feature, so that good nonlinear feature fitting can be performed on the multi-dimensional feature data set.
  • feature A has three attributes (A1, A2, A3)
  • feature B has two attributes (B1, B2)
  • the attributes of feature A are used to cross the features
  • 6 new features A1, B2), (A2, B2), (A3, B2), (B1, A1), (B1, A2) and (B1, A3).
  • the feature cross processing when performing feature cross processing, can be regarded as a logical AND operation of data.
  • the original features are first processed into bins, and then the binning results are processed.
  • By performing feature intersection better intersection data features can be obtained, thus greatly simplifying the calculation amount.
  • feature crossover processing can be performed by calculating a Cartesian product.
  • the Cartesian product is any two sets A and B. If the first member of the sequence pair is an element of A, and the second member is an element of B. All such even sets are called the Cartesian product or direct product of sets A and B, denoted as A ⁇ B.
  • the features generated through the above method contribute to the prediction target, that is, if they have a higher linear or nonlinear correlation than the original features, they are adopted as new features, that is, cross features.
  • Step S320 Perform feature derivation processing on the original features and cross features to obtain derived features.
  • the so-called feature derivation process refers to the creation of new features through existing data.
  • Feature derivation is sometimes also called feature creation, feature extraction, etc.
  • One is to create new features based on the characteristics of the data set.
  • the feature derivation at this time is actually a type of unsupervised feature derivation, such as combining monthly charges (MonthlyCharges) and total charges (TotalCharges).
  • the columns are added to create a new column; in another case, the data set label is also taken into consideration to create a new feature.
  • feature derivation is actually supervised feature derivation.
  • Most of the time feature derivation specifically refers to unsupervised feature derivation, while supervised feature derivation we will call it target encoding.
  • Feature derivation includes but is not limited to feature combination, feature intersection, image feature generation, text feature generation, etc.
  • feature combination can be realized through the combination of four arithmetic operations between pairs of features, logical AND, or combination, polynomial construction, the difference between the feature itself and its mean, etc.
  • Step S330 Combine original features, cross features and derived features to obtain target features.
  • the original features, cross features and derived features are combined to form the target features.
  • the principle of constructing the target features is to add all the features that people think may be meaningful, so as to retain as much of the original data as possible.
  • the amount of information enables the subsequent prediction model to perform better.
  • the decision tree model will adaptively filter out features that are not meaningful to the final result based on the different information gains brought by the features of each dimension.
  • feature construction is performed on the original features of the target product resources to obtain the target features.
  • the target features have less noise in the amount of information, that is, the proportion of effective features increases.
  • the target is determined based on the target features.
  • the predicted revenue probability value of product resources can be more accurate.
  • Step S220 Determine the predicted profit probability value of the target product resource based on the target characteristics.
  • the predicted profit probability value of the target product resource can be determined based on the constructed target characteristics. This predicted profit probability value can represent the probability of whether the target product resource is likely to make a profit in the next N days.
  • Figure 4 is a flow chart illustrating a data processing method for product resources according to an exemplary embodiment.
  • the data processing method for product resources includes steps S410 to S460. The details are as follows:
  • Step S410 Extract original features from the original data of the target product resources, and perform feature construction processing based on the original features to obtain the target features.
  • step S410 is consistent with the introduction of the above step S210, and will not be described again.
  • Step S420 Input the target features into the trained prediction model to obtain the predicted profit probability values of the target product resources in multiple dimensions.
  • the target features are constructed, they are input into the trained prediction model to obtain predicted profit probability values in multiple dimensions.
  • the predicted profit probabilities in multiple dimensions can be It includes the profit expectation of selling the target product resource on any day in the next N days, the profit expectation of selling the target product resource after a specified period of time, and the profit expectation of the target product resource under the most ideal conditions.
  • the prediction model will take the OR relationship between these three types of revenue expectations and obtain the predicted revenue probability value of whether the target product resources are likely to be profitable in the above three situations.
  • the above prediction model can be trained based on a decision tree model or a model such as XGBoost (eXtreme Gradient Boosting).
  • the decision tree model is an instance-based inductive learning method that can extract a tree-type classification model from given unordered training samples.
  • Each non-leaf node in the tree records which feature is used to judge the category, and each leaf node represents the final category judged.
  • a classified path rule is formed from the root node to each leaf node. When testing a new sample, you only need to start from the root node, test at each branch node, recursively enter the subtree along the corresponding branch and test again, until you reach the leaf node.
  • the category represented by the leaf node is is the predicted category of the current test sample.
  • the predicted profit probability value Por_val is determined as:
  • ⁇ represents the preset probability factor.
  • T regression trees will be constructed in the XGBoost model.
  • the t-th regression tree When the t-th regression tree is constructed, the residuals generated by the classification regression of the training samples from the first t-1 regression trees need to be fitted. Each time the fitting generates a new regression tree, all possible regression trees are traversed and the regression tree that minimizes the objective function value is selected. However, this is difficult to achieve in practice, so the steps need to be decomposed.
  • constructing a new regression tree only one branch is generated at a time and the best branch is selected. If the objective function value (cost) of generating a branch is greater than when it is not generated or the improvement effect is not obvious, then the branch will be abandoned.
  • Step S430 Determine the trend of the target product resource based on the predicted profit probability value of the target product resource and the predicted profit probability value of other product resources.
  • step S430 is consistent with the description of the later-described step S230, and will not be described here.
  • Step S440 Obtain the target decision path from multiple optional decision paths of the target product resource contained in the prediction model.
  • the prediction model is trained based on the decision tree model.
  • the decision tree model starts from the root node, tests a certain feature of the product resource, and allocates the product resource to its child nodes according to the test results. , at this time, each child node corresponds to a value of the feature, and the product resources are tested and allocated recursively until reaching the leaf node, and finally the product resources are divided into the classes of the leaf node. Therefore, multiple decision paths will be formed in the prediction model. After the target product resources are input into the prediction model, they will go through a decision path and reach the corresponding leaf node. This decision path is the target decision path.
  • Step S450 Determine the original interpretation information of the target product resource based on the node attributes on the target decision path.
  • the decision path includes multiple nodes, and these nodes have corresponding node attributes. These node attributes determine how subsequent nodes are split. According to the node attributes of the target decision path, the judgment of obtaining the predicted profit probability value can be determined. Conditions are summarized and processed to obtain the original interpretation information of the target product resources. The original interpretation information is a comprehensive interpretation of the target product resources.
  • Step S460 Obtain the target interpretation information of the target product resource based on the trend and the original interpretation information of the target product resource.
  • step S460 is consistent with the description of the later-described step S240, and will not be described here.
  • the technical solution provided in this embodiment obtains the predicted revenue probability value through the trained prediction model.
  • the prediction model can closely connect various data of the target product resources, and then provide the corresponding predicted revenue probability value.
  • the predicted profit probability value can be quickly calculated without relying on manual experience.
  • step S440 after obtaining the target decision path from the multiple optional decision paths of the target product resource contained in the prediction model, the data processing method of the product resource also includes step S510 -S520, detailed introduction is as follows:
  • Step S510 Match the target decision path of the target product resource with the historical decision path of each historical product resource.
  • the historical product resources are processed by the prediction model, so that the target decision path consistent with step S440 can be obtained.
  • the target decision path of the historical product resource is the corresponding historical decision path.
  • the target decision path of the target product resource is The decision path is matched with the historical decision path of historical product resources to determine whether the target decision path and the historical decision path are the same.
  • Step S520 Obtain the historical product resources corresponding to the historical decision paths that match the target decision path as reference product resources for the target product resources, so as to determine relevant decisions on the target product resources based on the information of the reference product resources.
  • the historical product resources corresponding to the historical decision path that are the same as the target decision path are used as the reference product resources of the target product resources.
  • step S420 before step S420 inputs the target features into the trained prediction model to obtain the predicted profit probability value of the product resource in multiple dimensions, the method also includes steps S610-S620. , the details are as follows:
  • Step S610 Obtain training sample data of sample product resources.
  • the sample product resources include multiple, and the training sample data of the sample product resources is obtained.
  • the training sample data of the sample product resources is equivalent to the original data of the target product resources.
  • a sufficient amount of sample product resources can be prepared for training, and all sample product resources can be input into the decision tree model for training; the sample product resources can also be divided into training groups and verification groups, using the training group After training the decision tree model, use the verification group to verify whether the performance of the trained decision tree model meets the requirements. If it does not meet the requirements, retrain.
  • Step S620 Obtain the flow distribution characteristics and increase characteristics of the sample product resources based on the training sample data.
  • the training sample data is subjected to the feature construction processing as mentioned above to obtain the corresponding flow direction distribution characteristics.
  • Which feature construction processing is used for the training sample data during training will be used for the original features of the target product resources.
  • the training sample data is actual data, so the growth characteristics can be obtained based on the training sample data.
  • the training sample data includes first data characterizing the flow direction of the sample product resources, second data characterizing the distribution of the sample product resources, and third data characterizing the rise and fall of the sample product resources; see Figure 7, step S620 Based on the training sample data, the flow distribution characteristics and growth characteristics of the sample product resources are obtained, including steps S710-S730, which are described in detail as follows:
  • Step S710 Perform feature construction processing based on the first data, the second data, and the third data to obtain the flow distribution characteristics of the sample product resources.
  • the first data, the second data and the third data are processed through the same feature structure as mentioned above, that is, the corresponding first features of the first data, the second data and the third data are subjected to feature intersection processing.
  • Obtain the second feature then perform feature derivation processing on the second feature and the first feature to obtain the third feature, and combine the first feature, the second feature and the third feature to obtain the flow direction distribution feature.
  • Step S720 Calculate the first income of the sample product resource in the average income dimension, the second income of the sample product resource in the random income dimension, and the third income of the sample product resource in the maximum income dimension based on the third data, and The growth characteristics of sample product resources are obtained based on the first income, second income and third income.
  • the third data represents the rise and fall of the sample product resources, that is, the third data records the price changes of the sample product resources within a preset time period.
  • the first income, the second income and the third income can be calculated based on the third data.
  • the first income is the income expectation of the sample product resources sold on any day in the next N days
  • the second income is the sample product resources after the specified time period.
  • the profit expectation of selling, and the third profit is the profit expectation that the sample product resources can obtain under the most ideal conditions.
  • the first profit is obtained by assuming that the sample product resources are purchased on T 0 day , and the sample product resources are sold on T 1 , T 2 , ..., T N days respectively in the next N days.
  • the different rates of return P 1 , P 2 ,..., P N are calculated based on the daily rate of return.
  • the first income can be obtained by the formula represents, among which, FAP represents the first profit, N represents the next N days, N is an integer greater than or equal to 1, i represents the i-th day, i is an integer less than or equal to N, and Pi is the sample product resource sold on the i-th day. income.
  • the second return is the return rate P N obtained by assuming that the sample product resources are purchased on day T 0 and the sample product resources are sold on day T N .
  • Step S730 Input the flow distribution characteristics and increase characteristics into the decision tree model for iterative training to obtain a prediction model.
  • the flow distribution characteristics and growth characteristics of the sample product resources are input into the decision tree model for iterative training, so that the trained prediction model can predict whether there will be profits in the next N days based on the target characteristics of the target product resources.
  • the prediction model after training the decision tree model to obtain the prediction model, can be updated according to the sliding window of weekly, monthly, yearly and other time periods, so that the prediction model after each update can be more adaptable to each time period. Forecast of product resources.
  • Step S230 Determine the trend of the target product resource based on the predicted profit probability value of the target product resource and the predicted profit probability value of other product resources.
  • the predicted profit probability value output by the prediction model according to the target characteristics is a numerical value, and this predicted profit probability value is standardized.
  • the distribution of predicted profit probability values corresponding to other product resources is considered, and the target product resources are The predicted profit probability value is classified into three trends: "up”, “no obvious trend” and "down”.
  • Step S240 Obtain the target interpretation information of the target product resource based on the trend and the original interpretation information of the target product resource.
  • the target product resources have corresponding original interpretation information.
  • the original interpretation information is obtained based on the original data.
  • the trend is inserted into the corresponding position of the original interpretation information, so that the complete target interpretation information is displayed on the target On the interpretation page of product resources, if the target interpretation information is: Recently, the maximum profit spread of chips is trend.
  • feature construction is performed on the original features of the target product resources to obtain the target features.
  • the target features have more information. Determining the predicted revenue probability value of the target product resources based on the target features can be more accurate.
  • the trend is also obtained based on the preset profit probability value, and the target interpretation information is obtained based on the trend trend and the original interpretation information.
  • the user understands the future direction of the target product resources by viewing the target interpretation information, and makes corresponding adjustments to the target product resources.
  • step S240 based on the trend and the original interpretation information of the target product resource, the target interpretation information of the target product resource is obtained, including step S810 and step S820.
  • the details are as follows:
  • Step S810 Process the original interpretation information through a preset analysis description algorithm to obtain processed original interpretation information.
  • an analysis description algorithm is preset.
  • the analysis description algorithm is based on Support Vector Regression (SVR).
  • SVR Support Vector Regression
  • the support vector regression model is a machine learning model for regression evolved from the support vector machine. , because this model has good prediction performance, and the derivation process of this model is to solve the solution of the Lagrangian dual equation on a convex set, which is a quadratic optimization problem, and the solution obtained by using this model is the global optimal solution. .
  • the original interpretation information processed by the analysis and description algorithm can describe the target product resources more smoothly and accurately.
  • Step S820 Combine the trend and the processed original interpretation information to obtain the target interpretation information of the target product resource.
  • the trend trend is inserted into a fixed position of the processed original interpretation information to form a complete sentence, and the target interpretation information of the target product resource can be obtained.
  • Figure 9 is a flow chart illustrating a data processing method for product resources according to an exemplary embodiment.
  • the data processing method for product resources includes steps S910 to step S9130. Detailed introduction as follows;
  • Step S910 based on the third data of the sample data, calculate the first income of the sample product resources in the average income dimension, the second income of the sample product resources in the random income dimension, and the third income of the sample product resources sold in the maximum income dimension. income, and obtain the growth characteristics of the sample product resources based on the first income, second income and third income.
  • the first income, the second income and the third income are calculated according to the third data of the sample data.
  • the first income, the second income and the third income have all been described above. This description is not repeated.
  • Step S920 Perform feature construction processing based on the first data, second data and third data of the sample product resources to obtain the flow distribution characteristics of the sample product resources.
  • feature intersection processing is performed on the first features corresponding to the first data, second data and third data to obtain the second feature, and then feature derivation processing is performed on the second feature and the first feature to obtain the third feature.
  • feature derivation processing is performed on the second feature and the first feature to obtain the third feature.
  • Step S930 Input the flow distribution characteristics and increase characteristics into the decision tree model for iterative training to obtain a prediction model.
  • the flow distribution characteristics and increase characteristics of the sample product resources are input into the decision tree model for iterative training, so that the trained prediction model can obtain whether the target product resource will be in the next N days based on the target characteristics of the target product resource.
  • the predicted profit probability value that is likely to make a profit.
  • Step S940 Extract original features from the original data of the target product resources, and perform feature construction processing based on the original features to obtain the target features.
  • the same feature construction process as described above is performed based on the original features to obtain the corresponding target features.
  • Step S950 Input the target features into the trained prediction model to obtain the predicted profit probability values of the target product resources in multiple dimensions.
  • the target features are input into the prediction model trained based on the decision tree model for calculation, and the corresponding predicted profit probability value is obtained.
  • Step S960 Determine the trend of the target product resource based on the predicted profit probability value of the target product resource and the predicted profit probability value of other product resources.
  • the predicted profit probability value output by the prediction model according to the target characteristics is a numerical value.
  • the predicted profit probability value of the target product resource is calculated based on the distribution of predicted profit probability values corresponding to other product resources. It is classified into three trends: "up”, “no obvious trend” and "down”.
  • Step S970 Obtain the target decision path from multiple optional decision paths of the target product resource contained in the prediction model.
  • the decision tree model has multiple decision paths, and each decision path has a leaf node.
  • the target product resources will eventually be classified into the correct leaf node, and the decision path corresponding to the correct leaf node is the target decision path of the target product resources.
  • Step S980 Determine the original interpretation information of the target product resource based on the node attributes on the target decision path.
  • each node in the decision tree model has corresponding node attributes. These node attributes determine how subsequent nodes are split. According to the node attributes of the target decision path, the judgment conditions for obtaining the predicted profit probability value can be determined. For this Perform inductive processing to obtain the original interpretation information of the target product resources, which is a comprehensive interpretation of the target product resources.
  • Step S990 Process the original interpretation information through a preset analysis description algorithm to obtain processed original interpretation information.
  • the original interpretation information processed by the analysis and description algorithm can describe the target product resources more smoothly and accurately.
  • Step S9100 Combine the trend and the processed original interpretation information to obtain the target interpretation information of the target product resource.
  • the processed original interpretation information has a vacancy specifically for placing the trend. Directly inserting the trend into the original interpretation information can form a complete sentence, which is the target interpretation information. Users can directly view the target interpretation information to learn whether profits are possible in the next N days, and then make relevant decisions about target product resources.
  • Step S9110 Match the target decision path of the target product resource with the historical decision path of each historical product resource.
  • the target decision path of the target product resource is matched with the historical decision path of historical product resources.
  • the historical decision path of the historical product resource is matched with the target product.
  • the resource-eating target decision path is obtained in the same way.
  • Step S9120 Obtain the historical product resources corresponding to the historical decision paths that match the target decision path as reference product resources for the target product resources, so as to determine relevant decisions on the target product resources based on the information of the reference product resources.
  • the target decision path matches the historical decision path, which means that the two go through the same path in the decision tree model, and the historical product resources corresponding to the historical decision path that matches the target decision path are directly used as the target product resources.
  • Reference product resources, the historical decision-making path of reference product resources are also used to determine whether it is possible to make a profit in the next N days, but based on the historical decision-making path of reference product resources, there are real corresponding data for the next N days. Therefore, the user Decisions can be made by viewing the real corresponding data of historical product resources in the next N days.
  • Step S9130 Display the target interpretation information and reference product resources as output results on the interpretation page of the target product resource.
  • the target interpretation information and the reference product resources are displayed to the user together, and the user can intuitively view the final results.
  • a jump button for the reference product resources can also be set on the interpretation page, so that the user can view the result from the interpretation page. Jump to the corresponding page of the reference product resources to observe the reference product resources more comprehensively.
  • the technical solution provided by the embodiment of this application processes the target characteristics of the target product resources through the trained prediction model to obtain the predicted revenue probability value, and processes the predicted revenue probability value based on the predicted revenue probability values of other product resources to obtain
  • the trend is inserted into the original interpretation information obtained from the target decision path to form the target interpretation information. It can rely on the information of the target product resources and rely on the algorithm to obtain the target interpretation information, which improves the conversion rate of information and At the same time as the reception rate, the amount of effective information is increased.
  • this embodiment also provides a reference product for target product resources. Resources, users can interpret information based on goals and refer to product resources to make relevant decisions.
  • an exemplary embodiment of the present application provides a data processing device for product resources, including:
  • the extraction module 1010 is configured to extract original features from the original data of the target product resources, and perform feature construction processing based on the original features to obtain the target features;
  • the first determination module 1020 is configured to determine the predicted revenue probability value of the target product resource based on the target characteristics
  • the second determination module 1030 is configured to determine the trend of the target product resource based on the predicted profit probability value of the target product resource and the predicted profit probability value of other product resources;
  • the target interpretation information module 1040 is configured to obtain the target interpretation information of the target product resources based on the trend and the original interpretation information of the target product resources.
  • the number of original features is multiple;
  • the extraction module 1010 includes: a first processing sub-module configured to perform feature intersection processing on multiple original features to obtain intersection features; a second processing sub-module, It is configured to perform feature derivation processing on original features and cross features to obtain derived features;
  • the combination submodule is configured to combine original features, cross features and derived features to obtain target features.
  • the first determination module 1020 includes: an input submodule configured to input target features into the trained prediction model to obtain predicted revenue probability values of the target product resources in multiple dimensions; device It also includes: a first acquisition module configured to acquire a target decision path from multiple optional decision paths of target product resources contained in the prediction model; a third determination module configured to determine the target product based on node attributes on the target decision path Original interpretation information of the resource.
  • the device further includes: a matching module configured to match the target decision path of the target product resource with the historical decision path of each historical product resource; a second acquisition module configured to obtain the target decision path corresponding to the target decision path.
  • the historical product resources corresponding to the matched historical decision paths are used as reference product resources for the target product resources to determine relevant decisions on the target product resources based on the information of the reference product resources.
  • the device further includes: a third acquisition module configured to acquire training sample data of the sample product resources; a feature module configured to obtain the flow distribution characteristics and increase characteristics of the sample product resources based on the training sample data;
  • the training module is configured to input flow distribution characteristics and increase characteristics into the decision tree model for iterative training to obtain a prediction model.
  • the training sample data includes first data characterizing the flow of sample product resources, second data characterizing the distribution of sample product resources, and third data characterizing the rise and fall of sample product resources;
  • the feature module includes: feature construction The sub-module is configured to perform feature construction processing based on the first data, the second data and the third data to obtain the flow distribution characteristics of the sample product resources; and the calculation sub-module is configured to calculate the average income of the sample product resources based on the third data.
  • the first income in the dimension, the second income in the random income dimension of the sample product resources, and the third income from the sale of the sample product resources in the maximum income dimension, and the sample is obtained based on the first income, the second income and the third income. Growth characteristics of product resources.
  • the target interpretation information module 1040 includes: a third processing sub-module configured to process the original interpretation information through a preset analysis description algorithm to obtain processed original interpretation information; a merging sub-module, It is configured to merge the trend and the processed original interpretation information to obtain the target interpretation information of the target product resources.
  • Embodiments of the present application also provide an electronic device, including: one or more processors; a storage device for storing one or more programs. When the one or more programs are processed by the one or more When the processor is executed, the electronic device is caused to implement the data processing method of product resources provided in the above embodiments.
  • FIG. 11 shows a schematic structural diagram of a computer system suitable for implementing an electronic device according to an embodiment of the present application.
  • the computer system 1100 includes a central processing unit (Central Processing Unit, CPU) 1101, which can be loaded into a random accessory according to a program stored in a read-only memory (Read-Only Memory, ROM) 1102 or from a storage part 1108. Access the program in the memory (Random Access Memory, RAM) 1103 to perform various appropriate actions and processing, such as performing the method described in the above embodiment. In RAM 1103, various programs and data required for system operation are also stored.
  • CPU 1101, ROM 1102 and RAM 1103 are connected to each other through bus 1104.
  • An input/output (I/O) interface 1105 is also connected to bus 1104.
  • the following components are connected to the I/O interface 1105: an input part 1106 including a keyboard, a mouse, etc.; an output part 1107 including a cathode ray tube (Cathode Ray Tube, CRT), a liquid crystal display (Liquid Crystal Display, LCD), etc., and a speaker, etc. ; a storage part 1108 including a hard disk, etc.; and a communication part 1109 including a network interface card such as a LAN (Local Area Network) card, a modem, etc.
  • the communication section 1109 performs communication processing via a network such as the Internet.
  • Driver 1110 is also connected to I/O interface 1105 as needed.
  • Removable media 1111 such as magnetic disks, optical disks, magneto-optical disks, semiconductor memories, etc., are installed on the drive 1110 as needed, so that a computer program read therefrom is installed into the storage portion 1108 as needed.
  • the process described above with reference to the flowchart may be implemented as a computer software program.
  • embodiments of the present application include a computer program product including a computer program carried on a computer-readable medium, the computer program including a computer program for performing the method shown in the flowchart.
  • the computer program may be downloaded and installed from the network via communication portion 1109 and/or installed from removable media 1111 .
  • the central processing unit CPU
  • various functions defined in the system of the present application are executed.
  • the computer-readable medium shown in the embodiments of the present application may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the above two.
  • the computer-readable storage medium may be, for example, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof.
  • Computer readable storage media may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard drive, random access memory (RAM), read only memory (ROM), removable Programmable Read-Only Memory (Erasable Programmable Read Only Memory, EPROM), flash memory, optical fiber, portable compact disk read-only memory (Compact Disc Read-Only Memory, CD-ROM), optical storage device, magnetic storage device, or any of the above suitable The combination.
  • a computer-readable storage medium may be any tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device.
  • a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, in which a computer-readable computer program is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above.
  • a computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device .
  • Computer programs embodied on computer-readable media may be transmitted using any suitable medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.
  • each block in the flow chart or block diagram may represent a module, program segment, or part of the code.
  • the above-mentioned module, program segment, or part of the code includes one or more executable components for implementing the specified logical function. instruction.
  • the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown one after another may actually execute substantially in parallel, or they may sometimes execute in the reverse order, depending on the functionality involved.
  • each block in the block diagram or flowchart illustration, and combinations of blocks in the block diagram or flowchart illustration can be implemented by special purpose hardware-based systems that perform the specified functions or operations, or may be implemented by special purpose hardware-based systems that perform the specified functions or operations. Achieved by a combination of specialized hardware and computer instructions.
  • the units involved in the embodiments of this application can be implemented in software or hardware, and the described units can also be provided in a processor. Among them, the names of these units do not constitute a limitation on the unit itself under certain circumstances.
  • Another aspect of the present application also provides a computer-readable storage medium on which a computer program is stored.
  • a computer program When the computer program is executed by a processor, the method as described above is implemented.
  • the computer-readable storage medium may be included in the electronic device described in the above embodiments, or may exist separately without being assembled into the electronic device.
  • Another aspect of the present application also provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium.
  • the processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the methods provided in the above embodiments.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Data Mining & Analysis (AREA)
  • Marketing (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Technology Law (AREA)
  • Game Theory and Decision Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Computation (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

Sont divulgués dans la présente demande un procédé et un appareil de traitement de données de ressource produit, un dispositif électronique et un support de stockage. Le procédé comprend : l'extraction de caractéristiques d'origine à partir de données d'origine de ressources produit cibles et la réalisation d'un traitement de construction de caractéristiques sur la base des caractéristiques d'origine pour obtenir des caractéristiques cibles ; la détermination d'une valeur de probabilité de revenu prédit des ressources produit cibles sur la base des caractéristiques cibles ; la détermination d'une tendance des ressources produit cibles sur la base de la valeur de probabilité de revenu prédit des ressources produit cible et de la valeur de probabilité de revenu prédit d'autres ressources produit ; et l'obtention d'informations d'interprétation cibles des ressources produit cibles sur la base de la tendance et des informations d'interprétation d'origine des ressources produit cibles.
PCT/CN2022/140756 2022-04-02 2022-12-21 Procédé et appareil de traitement de données de ressource produit, dispositif électronique et support de stockage WO2023185125A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210353762.4A CN114897607A (zh) 2022-04-02 2022-04-02 产品资源的数据处理方法及装置、电子设备、存储介质
CN202210353762.4 2022-04-02

Publications (1)

Publication Number Publication Date
WO2023185125A1 true WO2023185125A1 (fr) 2023-10-05

Family

ID=82715612

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/140756 WO2023185125A1 (fr) 2022-04-02 2022-12-21 Procédé et appareil de traitement de données de ressource produit, dispositif électronique et support de stockage

Country Status (2)

Country Link
CN (1) CN114897607A (fr)
WO (1) WO2023185125A1 (fr)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114897607A (zh) * 2022-04-02 2022-08-12 富途网络科技(深圳)有限公司 产品资源的数据处理方法及装置、电子设备、存储介质

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111178498A (zh) * 2019-12-09 2020-05-19 北京邮电大学 一种股票波动预测方法及装置
CN111476418A (zh) * 2020-04-08 2020-07-31 智谭(浙江)信息科技有限公司 基于深度神经网络和多任务学习的股票数据波动预测方法
CN111489259A (zh) * 2020-04-16 2020-08-04 江苏知诺智能科技有限公司 一种基于深度学习的股票市场风险预测智能实现方法
CN111563821A (zh) * 2020-05-12 2020-08-21 电子科技大学 一种基于支持向量机量化投资的金融股票涨跌预测方法
CN112365075A (zh) * 2020-11-19 2021-02-12 中国科学院深圳先进技术研究院 一种股票价格走势预测方法、系统、终端以及存储介质
CN113095484A (zh) * 2021-04-07 2021-07-09 华南师范大学 基于lstm神经网络的股票价格预测方法
CN113919944A (zh) * 2021-09-07 2022-01-11 暨南大学 基于强化学习算法与时间序列模型的股票交易方法及系统
CN114091789A (zh) * 2021-12-15 2022-02-25 中国平安人寿保险股份有限公司 项目数据波动预测方法、装置、计算机设备和存储介质
CN114897607A (zh) * 2022-04-02 2022-08-12 富途网络科技(深圳)有限公司 产品资源的数据处理方法及装置、电子设备、存储介质

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111178498A (zh) * 2019-12-09 2020-05-19 北京邮电大学 一种股票波动预测方法及装置
CN111476418A (zh) * 2020-04-08 2020-07-31 智谭(浙江)信息科技有限公司 基于深度神经网络和多任务学习的股票数据波动预测方法
CN111489259A (zh) * 2020-04-16 2020-08-04 江苏知诺智能科技有限公司 一种基于深度学习的股票市场风险预测智能实现方法
CN111563821A (zh) * 2020-05-12 2020-08-21 电子科技大学 一种基于支持向量机量化投资的金融股票涨跌预测方法
CN112365075A (zh) * 2020-11-19 2021-02-12 中国科学院深圳先进技术研究院 一种股票价格走势预测方法、系统、终端以及存储介质
CN113095484A (zh) * 2021-04-07 2021-07-09 华南师范大学 基于lstm神经网络的股票价格预测方法
CN113919944A (zh) * 2021-09-07 2022-01-11 暨南大学 基于强化学习算法与时间序列模型的股票交易方法及系统
CN114091789A (zh) * 2021-12-15 2022-02-25 中国平安人寿保险股份有限公司 项目数据波动预测方法、装置、计算机设备和存储介质
CN114897607A (zh) * 2022-04-02 2022-08-12 富途网络科技(深圳)有限公司 产品资源的数据处理方法及装置、电子设备、存储介质

Also Published As

Publication number Publication date
CN114897607A (zh) 2022-08-12

Similar Documents

Publication Publication Date Title
WO2020224405A1 (fr) Procédé et appareil de traitement d'image, support lisible par ordinateur et dispositif électronique
US11270375B1 (en) Method and system for aggregating personal financial data to predict consumer financial health
WO2020088058A1 (fr) Procédé et dispositif de génération d'informations
CN116127020A (zh) 生成式大语言模型训练方法以及基于模型的搜索方法
US11775504B2 (en) Computer estimations based on statistical tree structures
CN110909222A (zh) 基于聚类的用户画像建立方法、装置、介质及电子设备
CN109034199B (zh) 数据处理方法及装置、存储介质和电子设备
WO2023185125A1 (fr) Procédé et appareil de traitement de données de ressource produit, dispositif électronique et support de stockage
CN112241920A (zh) 基于图神经网络的投融资机构评估方法、系统及设备
WO2024051707A1 (fr) Procédé et appareil d'entraînement de modèle de recommandation, procédé et appareil de recommandation de ressources
US11645563B2 (en) Data filtering with fuzzy attribute association
CN112446777B (zh) 一种信用评估方法、装置、设备及存储介质
CN116401372A (zh) 知识图谱表示学习方法、装置、电子设备及可读存储介质
US11593740B1 (en) Computing system for automated evaluation of process workflows
CN112883256B (zh) 多任务处理方法、装置、电子设备及存储介质
WO2022271369A1 (fr) Apprentissage d'un modèle de liaison d'objets
CN114139059A (zh) 资源推荐模型的训练方法、资源推荐方法及装置
CN112307334A (zh) 信息推荐方法、信息推荐装置、存储介质与电子设备
CN113342998B (zh) 一种多媒体资源推荐方法、装置、电子设备及存储介质
CN114969543B (zh) 推广方法、系统、电子设备和存储介质
KR102449831B1 (ko) 신규 텍스트에 대한 정보를 제공하는 전자 장치, 신규 텍스트를 확인하는 서버 및 그 동작 방법
US20240121119A1 (en) Method and Apparatus for Classifying Blockchain Address
CN118227775A (zh) 咨询数据处理的方法、装置、电子设备和存储介质
CN118152759A (zh) 数据处理方法及装置、设备、存储介质和程序产品
CN117033757A (zh) 信息选取方法、装置、电子设备及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22934940

Country of ref document: EP

Kind code of ref document: A1