Feature verification task processing method and related product

Abstract

The invention provides a feature verification task processing method and device, equipment, a storage medium and a program product. The method comprises the steps of obtaining feature verification task description information input by a target user; generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information; a feature verification task is executed based on the feature verification task statement sequence, the feature verification task is used for executing each feature verification sub-task according to a target verification task process, and in the process of executing each feature verification sub-task, the training sample set is used as sample data, a target model corresponding to the target model parameter is used as a machine learning model, and the machine learning model is used for learning the feature verification sub-task. And verifying the feature evaluation index of the to-be-verified feature in each feature verification sub-task. According to the method, the feature evaluation process of the training sample set, the machine learning model, the model parameters and the at least one feature evaluation index can be customized without manually compiling codes, the evaluation process is simplified, and the evaluation time is saved.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the technical field of machine learning, to a method for processing tasks to be verified and related products, and specifically to a method, apparatus, device, storage medium and program product for processing feature verification tasks. Background Art

[0002] In the field of machine learning, feature mining is a commonly used method to enhance the prediction accuracy of models. At present, the mined features can generally be verified by calculating various evaluation indicators of the features. Feature evaluation indicators can be, for example, IV value of the feature, missing rate, equivalent rate, model gain, etc.

[0003] However, currently, verification of feature evaluation indicators is usually implemented through manual programming verification, which results in a complex, time-consuming, and difficult to manage and control verification process. Summary of the invention

[0004] The embodiments of the present disclosure provide a method, device, electronic device, storage medium and computer program product for processing a feature verification task.

[0005] In a first aspect, an embodiment of the present disclosure provides a method for processing a feature verification task, the method comprising:

[0006] Acquire feature verification task description information input by a target user, wherein the feature verification task description information includes training sample set description information, target model description information, target model parameter description information, feature description information to be verified, and target verification task process description information, wherein the target verification task process is used to indicate each feature verification subtask included in the feature verification task and the sequence between each feature verification subtask, wherein the feature verification subtask is used to determine at least one feature evaluation indicator;

[0007] Generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, wherein the feature verification task statement sequence is a statement sequence based on a logical language;

[0008] The feature verification task is performed based on the feature verification task statement sequence, wherein the feature verification task is used to execute each feature verification subtask according to the target verification task process, and in the process of executing each feature verification subtask, the training sample set is used as sample data, and the target model corresponding to the target model parameters is used as the machine learning model to verify the feature evaluation index of the feature to be verified in each of the feature verification subtasks.

[0009] In some optional implementations, before generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, the method further includes:

[0010] Acquire feature verification task scheduling strategy information input by the target user, wherein the feature verification task scheduling strategy information is used to indicate the execution time and / or execution order of feature verification subtasks in the feature verification task; and

[0011] The step of generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, and executing the feature verification task based on the feature verification task statement sequence, comprises:

[0012] A scheduling task is created and executed based on the feature verification task scheduling policy information, wherein the scheduling task is used to generate a feature verification subtask statement sequence for executing corresponding feature verification subtasks based on the feature verification task description information and the feature verification task scheduling policy information according to the execution time and / or execution order of each feature verification subtask indicated by the feature verification task scheduling policy information, and execute the corresponding feature verification subtask based on the feature verification subtask statement sequence.

[0013] In some optional implementations, before generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, the method further includes:

[0014] Acquire feature verification task log record configuration information input by the target user, wherein the feature verification task log record configuration information is used to indicate the log record configuration during the execution of a feature verification subtask in the feature verification task; and

[0015] Before executing the feature verification task based on the feature verification task statement sequence, the method further includes:

[0016] Distribute the logging configuration of each feature verification subtask to the corresponding feature verification subtask.

[0017] In some optional implementations, the feature verification task is also used to output a log record of the execution process of each feature verification subtask.

[0018] In some optional implementations, before generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, the method further includes:

[0019] Acquire the feature verification task trace description information input by the target user, wherein the feature verification task trace description information is used to indicate the code location and trace method that need to be recorded during the execution of the feature verification subtask in the feature verification task; and

[0020] The step of generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information includes:

[0021] A feature verification task statement sequence for executing the feature verification task is generated based on the feature verification task description information and the feature verification task trace description information.

[0022] In some optional implementations, after executing the feature verification task based on the feature verification task statement sequence, the method further includes:

[0023] For the output results of each feature verification subtask, a feature evaluation result report is generated.

[0024] In a second aspect, an embodiment of the present disclosure provides a feature verification task processing device, the device comprising:

[0025] A first acquisition module is configured to acquire feature verification task description information input by a target user, wherein the feature verification task description information includes training sample set description information, target model description information, target model parameter description information, to-be-verified feature description information, and target verification task process description information, wherein the target verification task process is used to indicate each feature verification subtask included in the feature verification task and the sequence between each feature verification subtask, wherein the feature verification subtask is used to determine at least one feature evaluation indicator;

[0026] A generating module, configured to generate a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, wherein the feature verification task statement sequence is a statement sequence based on a logical language;

[0027] An execution module is configured to execute the feature verification task based on the feature verification task statement sequence, wherein the feature verification task is used to execute each feature verification subtask according to the target verification task process, and in the process of executing each feature verification subtask, the training sample set is used as sample data, and the target model corresponding to the target model parameters is used as a machine learning model to verify the feature evaluation index of the feature to be verified in each of the feature verification subtasks.

[0028] In some optional embodiments, the device further comprises:

[0029] A second acquisition module is configured to acquire, before generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, feature verification task scheduling strategy information input by the target user, wherein the feature verification task scheduling strategy information is used to indicate the execution time and / or execution order of feature verification subtasks in the feature verification task; and

[0030] The generating module and the executing module are further configured to:

[0031] A scheduling task is created and executed based on the feature verification task scheduling policy information, wherein the scheduling task is used to generate a feature verification subtask statement sequence for executing corresponding feature verification subtasks based on the feature verification task description information and the feature verification task scheduling policy information according to the execution time and / or execution order of each feature verification subtask indicated by the feature verification task scheduling policy information, and execute the corresponding feature verification subtask based on the feature verification subtask statement sequence.

[0032] In some optional embodiments, the device further comprises:

[0033] A third acquisition module is configured to acquire the feature verification task log recording configuration information input by the target user before generating the feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, wherein the feature verification task log recording configuration information is used to indicate the log recording configuration during the execution of the feature verification subtask in the feature verification task; and

[0034] The configuration distribution module is configured to distribute the logging configuration of each feature verification subtask to the corresponding feature verification subtask before executing the feature verification task based on the feature verification task statement sequence.

[0035] In some optional implementations, the feature verification task is also used to output a log record of the execution process of each feature verification subtask.

[0036] In some optional embodiments, the device further comprises:

[0037] A fourth acquisition module is configured to obtain the feature verification task trace description information input by the target user before generating the feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, wherein the feature verification task trace description information is used to indicate the code position and trace method that need to be recorded during the execution of the feature verification subtask in the feature verification task; and

[0038] The generation module is further configured to generate a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information and the feature verification task trace description information.

[0039] In some optional embodiments, the device further comprises:

[0040] The result generation module is configured to generate a feature evaluation result report for the output result of each feature verification subtask after executing the feature verification task based on the feature verification task statement sequence.

[0041] In a third aspect, an embodiment of the present disclosure provides an electronic device, comprising: one or more processors; a storage device on which one or more programs are stored, and when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation manner in the first aspect.

[0042] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by one or more processors, implements the method described in any implementation manner in the first aspect.

[0043] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, including a computer program / instruction, which implements the method described in any implementation manner in the first aspect when the computer program / instruction is executed by a processor.

[0044] In order to solve the problem that the verification process of manually programming feature evaluation indicators is complicated, time-consuming, and difficult to manage and control, the embodiment of the present disclosure provides a feature verification task processing method, device, electronic device, storage medium and computer program product, by first obtaining feature verification task description information input by the target user, the feature verification task description information includes training sample set description information, target model description information, target model parameter description information, to-be-verified feature description information and target verification task process description information, the target verification task process is used to indicate the feature verification subtasks included in the feature verification task and the sequence between the feature verification subtasks, wherein the feature verification subtask is used to determine at least one feature evaluation indicator; then, based on the feature verification task description information, a feature verification task statement sequence for executing the feature verification task is generated, wherein the feature verification task statement sequence is a statement sequence based on a logical language; finally, the feature verification task is executed based on the feature verification task statement sequence, wherein the feature verification task is used to execute each feature verification subtask according to the target verification task process, and in the process of executing each feature verification subtask, the training sample set is used as sample data, and the target model corresponding to the target model parameter is used as a machine learning model to verify the feature evaluation indicator of the to-be-verified feature in each feature verification subtask. This makes it possible to implement a feature evaluation process that customizes training sample sets, machine learning models, model parameters, and at least one feature evaluation indicator without manually writing code, simplifying the evaluation process, saving evaluation time, and facilitating management and control. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Other features, objects and advantages of the present disclosure will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. In the drawings:

[0046] Figure 1 is an exemplary system architecture diagram in which an embodiment of the present disclosure may be applied;

[0047] Figure 2 is a flow chart of an embodiment of a method for processing a feature verification task according to the present disclosure;

[0048] Figure 3 It is a schematic diagram of the main structure between each sub-library (or functional module) in the feature evaluation library file disclosed in the present invention;

[0049] Figure 4 is a structural schematic diagram of an embodiment of a feature verification task processing device according to the present disclosure;

[0050] Figure 5 It is a structural diagram of a computer system of an electronic device suitable for implementing an embodiment of the present disclosure. DETAILED DESCRIPTION

[0051] The present disclosure is further described in detail below in conjunction with the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0052] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0053] Figure 1 An exemplary system architecture 100 is shown to which embodiments of the feature verification task processing method, apparatus, electronic device, and storage medium disclosed in the present invention can be applied.

[0054] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0055] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various client applications can be installed on terminal devices 101, 102, 103, such as feature evaluation applications, short video social applications, audio and video conferencing applications, web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0056] Terminal devices 101, 102, 103 can be hardware or software. When terminal devices 101, 102, 103 are hardware, they can be various electronic devices with information input devices (e.g., keyboard, mouse, touch screen, microphone, camera, etc.) and information output devices (e.g., display screen, speaker, etc.), including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III, Moving Picture Experts Compression Standard Audio Layer 3), MP4 (Moving Picture Experts Group Audio Layer IV, Moving Picture Experts Compression Standard Audio Layer 4) players, laptop computers and desktop computers, etc. When terminal devices 101, 102, 103 are software, they can be installed in the terminal devices listed above. It can be implemented as multiple software or software modules (for example, used to provide feature evaluation services), or it can be implemented as a single software or software module. No specific limitation is made here.

[0057] In some cases, the feature verification task processing method provided by the present disclosure may be executed by the terminal devices 101, 102, 103, and accordingly, the feature verification task processing apparatus may be disposed in the terminal devices 101, 102, 103. In this case, the system architecture 100 may not include the server 105.

[0058] In some cases, the feature verification task processing method provided by the present disclosure can be jointly executed by the terminal devices 101, 102, 103 and the server 105. For example, the step of "obtaining the feature verification task description information input by the target user" can be executed by the terminal devices 101, 102, 103, and the steps of "generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information" can be executed by the server 105. The present disclosure does not limit this. Accordingly, the feature verification task processing device can also be respectively set in the terminal devices 101, 102, 103 and the server 105.

[0059] In some cases, the feature verification task processing method provided in the present disclosure may be executed by the server 105 , and accordingly, the feature verification task processing device may also be set in the server 105 . In this case, the system architecture 100 may not include the terminal devices 101 , 102 , and 103 .

[0060] It should be noted that the server 105 can be hardware or software. When the server 105 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or it can be implemented as a single server. When the server 105 is software, it can be implemented as multiple software or software modules (for example, for providing distributed services), or it can be implemented as a single software or software module. No specific limitation is made here.

[0061] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to the implementation requirements.

[0062] Continue to refer Figure 2 , which shows a process 200 of an embodiment of a feature verification task processing method according to the present disclosure, the feature verification task processing method comprises the following steps:

[0063] Step 201: Obtain feature verification task description information input by a target user.

[0064] In this embodiment, the execution subject of the feature verification task processing method (for example Figure 1 The terminal device shown in the figure) can obtain the feature verification task description information input by the target user locally or remotely from other electronic devices connected to the above-mentioned execution subject network.

[0065] Here, the feature verification task description information may include training sample set description information, target model description information, target model parameter description information, to-be-verified feature description information, and target verification task process description information.

[0066] The training sample set description information is used to indicate the training sample set. For example, the training sample set description information may include but is not limited to the access address of the training sample set, the access method (such as the access username and password for accessing the database), etc. Optionally, the training sample set description information may also include a training sample condition, which is used to filter the data in the access address. The data that meets the training sample condition will be used as a training sample to form a training sample set. The training sample condition can be set according to the scheme. As an example, when the training sample set is a user information data set, the training sample condition can be that the user age in the user information is between 25 and 45 years old. The user information that meets the above conditions will be used as a training sample to form a training sample set.

[0067] The target model description information is used to indicate the target model. The target model can be any machine learning model. For example, the target model description information may include but is not limited to the address of the target model, the model file name, etc. Optionally, the target model description information may also include a target model condition, which is used to filter the machine learning model. The machine learning model that meets the target model condition will be used as the target model. The target model condition can be set according to the scheme. As an example, the target model condition can be a model including a multi-scale feature extraction network, or a neural network model including an attention mechanism, or a model with a number of model parameters less than a positive integer N, and so on.

[0068] The target model parameter description information is used to indicate various model parameters in the target model. For example, when the target model is a convolutional neural network, the target model parameters may include but are not limited to the number of network layers of the convolutional neural network, the convolution kernel parameters, weight parameters, bias parameters of the convolution layer, the pooling method of the pooling layer, etc. Optionally, the target model parameter description information may also include target model parameter conditions, which are used to constrain the model parameters of the target model.

[0069] The feature description information to be verified is used to indicate the features to be verified for feature extraction of the training samples in the training sample set. Here, the features are various data obtained after feature acquisition, calculation, and statistics of at least one data included in the training sample. For example, feature extraction of user information can obtain but is not limited to the following features:

[0070] (1) Basic information characteristics: account name, language preference;

[0071] (2) Behavioral characteristics: browsing history (web pages, applications, products), search history, purchase history, user behavior path in the application, click-through rate (CTR), conversion rate;

[0072] (3) Social network characteristics: frequency of social network use, number of relationships in the social network (friends, followers, fans), content generation in the social network (posting, commenting, sharing), and interaction in the social network (likes, reposts);

[0073] (4) User preference characteristics: product or service preference, brand loyalty, price sensitivity, and purchase frequency;

[0074] (5) Temporal characteristics: active time (the time period during which users are active), access interval (the time interval between two logins), seasonal patterns (the impact of holidays and seasons on behavior);

[0075] (6) Technology usage characteristics: device type (mobile phone, tablet, computer), operating system (iOS, Android, Windows), browser type, and application version;

[0076] (7) Content consumption characteristics: type of content read (news, blogs, videos), reading time, and content interaction (comments, ratings);

[0077] (8) User feedback characteristics: product evaluation, service evaluation, and customer support interaction;

[0078] (9) Transaction characteristics: transaction amount, transaction frequency, and payment method;

[0079] (10) User status characteristics: user registration status (new user, old user), user churn risk, user level or membership status;

[0080] (11) User profile features: user grouping or segmentation (group division based on behavior, preferences, etc.);

[0081] (12) Predictive features: Future behaviors predicted based on historical data (e.g., purchase probability, churn probability).

[0082] It should be noted that the information, data and signals involved in this disclosure are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data comply with the relevant laws, regulations and standards of relevant countries and regions.

[0083] The target verification task process description information is used to indicate the target verification task process, and the target verification task process is used to indicate the feature verification subtasks included in the feature verification task and the sequence between the feature verification subtasks. The feature verification subtasks are used to determine at least one feature evaluation indicator. For example, the target verification task process description information may include the evaluation process template identifier, storage address, access method, etc. corresponding to the target verification task process. Optionally, the target verification task process description information may also include a target verification task process condition, which is used to constrain the verification task process template. The verification task process template that meets the target verification task process condition will be used as the target verification task process.

[0084] Here, the feature evaluation index may include various feature evaluation indexes currently available or to be developed in the future, and the present disclosure does not specifically limit this. As an example, the feature evaluation index includes but is not limited to:

[0085] Missing rate, identical value rate, IV (Information Value) value, model gain, dispersion, stability, outlier ratio, interpretability, data distribution, correlation coefficient, mutual information and maximum information coefficient, chi-square test, information gain, Gini impurity, feature importance, permutation importance, cross-validation score (accuracy, F1 score), etc.

[0086] Here, the execution subject of step 201 may be installed with various interactive notebooks that support at least one programming language, such as Jupyter Notebook (the essence of Jupyter Notebook is a web application that is convenient for creating and sharing program documents, supporting real-time code, mathematical equations, visualization and markdown). The target user can input feature verification task description information by utilizing the input interface provided by the above-mentioned interactive notebook. For example, the target user can enter feature verification task description information in a cell of Jupyter Notebook. That is, the target user can enter feature verification task description information in a metaprogramming manner without entering the complete code. Furthermore, the execution subject of step 201 can also obtain the feature verification task description information input by the target user through the above-mentioned interactive notebook, and store the feature verification task description information input by the target user as various interactive notebook files. For example, it can be stored as a Jupyter Notebook file in .ipynb format.

[0087] In practice, multiple data sets, multiple machine learning models, multiple features, multiple model parameters, and multiple verification task flow templates can be pre-formulated and stored. The target user can select a training sample set from the above data sets, select a target model from multiple machine learning models, select features to be verified from multiple features, select target model parameters from multiple model parameters, and select a target verification task flow from multiple verification task flow templates through an interactive notebook.

[0088] Step 202: Generate a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information.

[0089] Here, the execution subject of the feature verification task processing method can generate a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information in various implementation methods. Among them, the feature verification task statement sequence is a statement sequence based on a logical language. As an example, the feature verification task statement sequence can be a source code (Source Code), an object code (Object Code), an executable code (Executable Code), a bytecode (Bytecode), an assembly code (Assembly Code), a machine code (Machine Code), a pseudocode (Pseudocode), a script code (Script Code), a library code (Library Code), etc., which are used to execute the feature verification task and embody the logic of the feature verification task.

[0090] Specifically, the execution subject of step 202 can generate a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information using a code conversion tool. Here, the code conversion tool can convert an interactive notebook file into a code file. For example, the command line tool nbconvert in Jupyter Notebook (nbconvert is a command line tool for Jupyter Notebook, which can convert Jupyter Notebook files in .ipynb format into a variety of other formats, such as HTML, LaTeX, PDF, Markdown, etc.). Then, the execution subject of step 202 can convert the interactive notebook file generated based on the feature verification task description information in step 201, such as a Jupyter Notebook file in .ipynb format, into a feature verification task statement sequence using a code conversion tool. For example, the feature verification task statement sequence can be an HTML, LaTeX or Markdown file.

[0091] In another embodiment, the step of generating a feature verification task statement sequence based on the feature verification task description information at least includes:

[0092] (1) Preprocessing: According to the feature verification task description information, an initial task statement sequence in the first format configured with the feature task description information is obtained, and the initial task statement sequence is pre-processed (such as clearing the corresponding output statement sequence).

[0093] (2) Conversion parameter setting: Select a target format from a plurality of second formats (such as HTML, LaTeX, PDF, Markdown, etc.) or select a target template from a plurality of statement sequence templates, wherein the target template is used to set the target format of the feature verification task statement sequence.

[0094] (3) Conversion processing (i.e., the process of generating a feature verification task statement sequence): Based on the selected conversion parameters, the statement sequence library of the target format is called to generate and output the feature verification task statements corresponding to each task statement in the initial task statement sequence. Furthermore, the generated and output task statements can form a feature verification task statement sequence in the order corresponding to the initial task statement sequence.

[0095] As an example, the feature verification task statement sequence may be a code, and the code conversion logic for converting the interactive notebook file (i.e., the interactive notebook file generated based on the feature verification task description information in step 201) into a code file (i.e., the feature verification task statement sequence) using nbconvert may be performed as follows:

[0096] Preprocessors: nbconvert can use preprocessors to process interactive notebooks before conversion. For example, it can execute the code in the notebook, clean up the output, or perform other forms of processing on the notebook contents.

[0097] Exporters: The core part of nbconvert is the exporter, which defines how to convert notebooks into different formats. Each exporter is responsible for generating output in a specific format.

[0098] Templates: nbconvert uses Jinja-based templates to format the output. Users can customize the templates to control the look and feel of the output.

[0099] Postprocessors: After the conversion is complete, nbconvert can use postprocessors to further process the output, for example, to optimize HTML / CSS or process additional resource files.

[0100] Command Line Interface (CLI): nbconvert interacts primarily through the command line interface, where users can specify the conversion format and other options through command line parameters.

[0101] Configuration files: nbconvert supports the use of configuration files to customize the conversion behavior, including specifying the default export format, modifying preprocessor settings, or customizing template paths.

[0102] Execute conversion: The user enters the corresponding command through the command line, such as jupyter nbconvert --to htmlnotebook.ipynb, to start the conversion process.

[0103] Output: The converted file will be saved in the current working directory, and if there are supporting files (such as images), they will be placed in a _files directory named after the notebook.

[0104] Step 203: Execute the feature verification task based on the feature verification task statement sequence.

[0105] Specifically, the execution subject of step 203 may also be installed with various interactive notebooks supporting at least one programming language, such as Jupyter Notebook. Since the interactive notebook has the function of executing code, the execution subject of step 203 may use the interactive notebook (such as Jupyter Notebook) to execute the feature verification task statement sequence generated in step 202.

[0106] Here, the feature verification task is used to execute each feature verification subtask according to the target verification task process. In the process of executing each feature verification subtask, the training sample set is used as the sample data, and the target model corresponding to the target model parameters is used as the machine learning model to verify the feature evaluation index of the feature to be verified in each feature verification subtask.

[0107] For example, the feature verification task T specified by the target user in step 201 includes feature verification subtasks T1 and T2 executed sequentially, the training sample set is the user information set UserInfo, the target model is the convolutional neural network N1, the network structure S and model parameters P of the convolutional neural network are specified in the target model parameters, and the features to be verified include the IV value of the age field in the user information and the feature IV value of the occupation field.

[0108] In step 202 , a feature verification task statement sequence may be generated based on the above information. In step 203 , based on the feature verification task statement sequence generated in step 202 , feature verification subtasks T1 and T2 are sequentially executed.

[0109] Among them, the feature verification subtask T1 is used to verify the feature IV value of the age field in the user information using the user information set UserInfo as sample data and the convolutional neural network N1 set according to the network structure S and model parameters P as the machine learning model.

[0110] The feature verification subtask T2 is used to verify the feature IV value of the occupation field in the user information using the user information set UserInfo as sample data and the convolutional neural network N1 set according to the network structure S and model parameters P as the machine learning model.

[0111] Furthermore, by first performing the feature verification subtask T1, the model gain of the age field in the user information is obtained. Then, by performing the feature verification subtask T2, the feature IV value of the occupation field in the user information is obtained. That is, through step 203, the following two feature evaluation metrics can be obtained: the model gain of the age field in the user information and the feature IV value of the occupation field.

[0112] As an example, in the user credit assessment scenario, the IV value is used to evaluate the effectiveness of a feature in distinguishing different categories (such as good customers and bad customers).

[0113] Generally speaking, the threshold of the IV value can be used as a reference for feature selection:

[0114] · IV < 0.02: The predictive ability of the feature is weak and can usually be ignored.

[0115] · 0.02 < IV < 0.1: The feature has a certain predictive ability, but may not be a key feature.

[0116] · IV > 0.3: The feature has a high predictive ability and is a key feature of the model.

[0117] Suppose that after step 203, the feature IV value of the age field is less than 0.02, while the feature IV value of the occupation is greater than 0.3. This indicates that the occupation field is more suitable as a feature for predicting whether a user is a good customer or a bad customer using the target model as a machine learning model, while the age field is not suitable as a feature for evaluating the quality of customers.

[0118] Furthermore, by understanding and applying the IV value, feature selection and model construction can be carried out more effectively.

[0119] In this process, no code needs to be written. Only the description information of the feature verification task needs to be specified, which is convenient and fast and simplifies the feature verification process.

[0120] Through steps 201 to 203, it is possible to input the description information of the feature verification task using an interactive notebook, and without manually writing code, it is possible to implement the feature evaluation process of customizing the training sample set, machine learning model, model parameters, and at least one feature evaluation metric, simplify the evaluation process, save evaluation time, and facilitate management and control.

[0121] In some alternative embodiments, before step 202, the above feature verification task processing method flow 200 may further include the following step 204:

[0122] Step 204, obtaining the feature verification task scheduling policy information input by the target user.

[0123] Here, the feature verification task scheduling strategy information is used to indicate the execution time and / or execution order of the feature verification subtasks in the feature verification task.

[0124] Specifically, the execution body of step 204 may be installed with various interactive notebooks that support at least one programming language, such as Jupyter Notebook. The target user may use the input interface provided by the interactive notebook to input feature verification task scheduling policy information. For example, the target user may enter feature verification task scheduling policy information in a cell of Jupyter Notebook. That is, the target user may enter feature verification task scheduling policy information in a metaprogramming manner without entering the complete code. Furthermore, the execution body of step 204 may also obtain the feature verification task scheduling policy information input by the target user through the above-mentioned interactive notebook, and store the feature verification task description information input by the target user in step 201 and the feature verification task scheduling policy information input by the target user in step 204 as various interactive notebook files. For example, it may be stored as a Jupyter Notebook file in .ipynb format.

[0125] Accordingly, step 202 and step 203, i.e., generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, and executing the feature verification task based on the feature verification task statement sequence, may be performed as follows:

[0126] Create and execute scheduling tasks based on feature verification task scheduling strategy information.

[0127] The scheduling task here is used to generate a feature verification subtask statement sequence for executing the corresponding feature verification subtask based on the feature verification task description information and the feature verification task scheduling strategy information according to the execution time and / or execution order of each feature verification subtask indicated by the feature verification task scheduling strategy information, and to execute the corresponding feature verification subtask based on the feature verification subtask statement sequence.

[0128] In another embodiment, the step of generating a feature verification task statement sequence for executing a corresponding feature verification subtask based on the feature verification task description information and the feature verification task scheduling strategy information at least includes:

[0129] (1) Preprocessing: According to the feature verification task description information and the feature verification task scheduling strategy information, an initial subtask statement sequence in the first format configured with the task description information of the corresponding feature verification subtask is obtained, and the initial subtask statement sequence is preset (such as clearing the corresponding output statement sequence).

[0130] (2) Conversion parameter setting: Select a target format from multiple second formats (such as HTML, LaTeX, PDF, Markdown, etc.) or select a target template from multiple statement sequence templates, where the target template is used to set the target format of the feature verification subtask statement sequence.

[0131] (3) Conversion processing (i.e., the process of generating a feature verification subtask statement sequence): Based on the selected conversion parameters, the statement sequence library of the target format is called to generate and output the feature verification subtask statements corresponding to each subtask statement in the initial subtask statement sequence. Furthermore, each subtask statement generated and output can form a feature verification subtask statement sequence of the corresponding feature verification subtask in the order corresponding to the initial subtask statement sequence.

[0132] Since the feature verification subtask statement sequence is generated and executed according to the execution time and / or execution order of each feature verification subtask indicated by the feature verification task scheduling strategy information, the execution time and / or execution order of each feature verification subtask is also consistent with that indicated by the feature verification task scheduling strategy information.

[0133] Specifically, the execution subject of step 202 and step 203 may also be installed with a task scheduling tool that supports background execution of asynchronous processing and scheduled tasks. For example, Celery (Celery is a distributed task queue used to execute asynchronous tasks in the background. It uses a message agent to receive task requests and then distributes tasks to working nodes for execution). Furthermore, the execution subject of step 202 and step 203 may use the task scheduling tool to create a scheduling task based on the feature verification task scheduling strategy information, and execute the above scheduling task in the background.

[0134] Through the above optional implementation, it is possible to schedule the feature verification subtasks in the feature verification task according to the time and order specified by the target user, thereby designing the execution time and order of each feature verification subtask according to business needs without the need for complex coding.

[0135] In some optional implementations, before step 202, the feature verification task processing method flow 200 may further include the following steps 205 and 206:

[0136] Step 205: Acquire the feature verification task log record configuration information input by the target user.

[0137] Here, the feature verification task log configuration information is used to indicate the log configuration during the execution of the feature verification subtask in the feature verification task. For example, the log configuration may include but is not limited to: log format, print interval, log output mode (for example, storage in a log file or display on a terminal device).

[0138] Specifically, the execution body of step 205 can be installed with various interactive notebooks that support at least one programming language, such as Jupyter Notebook. The target user can use the input interface provided by the interactive notebook to input the feature verification task log recording configuration information. For example, the target user can enter the feature verification task log recording configuration information in a cell of Jupyter Notebook. That is, the target user can enter the feature verification task log recording configuration information in a metaprogramming manner without entering the complete code. Furthermore, the execution body of step 205 can also obtain the feature verification task log recording configuration information input by the target user through the above-mentioned interactive notebook, and store the feature verification task description information input by the target user in step 201 and the feature verification task scheduling strategy information input by the target user in step 205 as various interactive notebook files. For example, it can be stored as a Jupyter Notebook file in .ipynb format.

[0139] Thus, before step 203, that is, before executing the feature verification task based on the feature verification task statement sequence, the above method flow 200 may further include the following step 206:

[0140] Step 206: Distribute the logging configuration of each feature verification subtask to the corresponding feature verification subtask.

[0141] Here, the execution subject of step 206 may be installed with various log management tools. For example, the logging module in Python (the logging module in Python is a powerful log management system that can provide logs of different levels and can output logs to different places). Furthermore, the execution subject of step 206 may first determine the logging configuration of each feature verification subtask based on the feature verification task logging configuration information, and then use the above-mentioned log management tool (for example, the logging module in Python) to distribute the logging configuration of each feature verification subtask to the corresponding feature verification subtask.

[0142] Accordingly, the feature verification task is also used to output a log record of the execution process of each feature verification subtask. Here, the output of the log record can be in various forms, which are specifically specified by the target user through the feature verification task log record configuration information. For example, a log record file can be generated for storage, or the log record can be output to a terminal device for presentation.

[0143] Through the optional implementation of the above step 205 and step 206, it is possible to implement a logging method for each feature verification subtask specified by the target user without complicated coding.

[0144] In some optional implementations, before generating a feature verification task statement sequence based on the feature verification task description information in step 202, the above method flow may further include the following step 207:

[0145] Step 207, obtaining the feature verification task trace description information input by the target user.

[0146] Here, the feature verification task trace description information is used to indicate the code location and trace method that need to be recorded during the execution of the feature verification subtask in the feature verification task.

[0147] Specifically, the execution body of step 207 may be installed with various interactive notebooks that support at least one programming language, such as Jupyter Notebook. The target user can use the input interface provided by the interactive notebook to input the feature verification task trace description information. For example, the target user can enter the feature verification task trace description information in a cell of Jupyter Notebook. That is, the target user can enter the feature verification task trace description information in a metaprogramming manner without entering the complete code. Furthermore, the execution body of step 207 can also obtain the feature verification task trace description information input by the target user through the above-mentioned interactive notebook, and store the feature verification task trace description information input by the target user in step 201 and the feature verification task trace description information input by the target user in step 207 as various interactive notebook files. For example, it can be stored as a Jupyter Notebook file in .ipynb format.

[0148] As an example, in Jupyter Notebook, the traces of the execution process can be achieved in the following ways:

[0149] (1) Automatic saving: Jupyter Notebook can be set to automatically save, so that historical work will be saved regularly even when the browser is closed.

[0150] (2) Use the %save magic command: This command can save variables in a specific notebook so that they can be restored in other notebooks. For example:

[0151] Python

[0152] %save my_notebook data

[0153] This will save the variable data in my_notebook.

[0154] (3) Use the %store magic command: This command can share variables between different notebooks. For example:

[0155] Python

[0156] %store data

[0157] It can then be restored using %store -r data in another notebook.

[0158] (4) Use the %capture magic command: This command can capture the output and store it in a variable for later use. For example:

[0159] Python

[0160] %% capture output

[0161] #Some code

[0162] output.show()

[0163] This captures the output of the code block and stores it in the output variable.

[0164] (4) Use the %history magic command: This command can display historical code and its running results to help review previous work.

[0165] (5) Export notebooks: Users can export notebooks to different formats at any time, such as HTML, Markdown, PDF, etc.

[0166] (6) Use a version control system: Put the .ipynb file into a version control system (such as Git) to track the historical changes of the notebook.

[0167] (7) Use the %%time magic command: This command can measure the execution time of a code block, which is very helpful for performance analysis.

[0168] Thus, step 202, generating a feature verification task statement sequence based on the feature verification task description information, can be performed as follows:

[0169] A feature verification task statement sequence is generated based on the feature verification task description information and the feature verification task trace description information.

[0170] Specifically, the executor of step 202 can use a code conversion tool, such as nbconvert, to convert the interactive notebook file generated in step 208, such as a .ipynb format file, which includes the feature verification task description information input by the target user and the feature verification task trace description information, into a feature verification task statement sequence.

[0171] Through the above optional implementation method, it is possible to record the execution process of each feature verification subtask in the feature verification task in a traceable manner specified by the target user, which is convenient for monitoring and analyzing the feature evaluation process without the need for complex coding.

[0172] In some optional implementations, after step 203, the feature verification task processing method flow 200 may further include the following step 208:

[0173] Step 208: Generate a feature evaluation result report for the output result of each feature verification subtask.

[0174] Here, the execution body of step 208 may be installed with a report generation tool, such as openpyxl (openpyxl is a Python library for reading and writing Excel files) and xlrd (xlrd is a Python library for reading Excel files (especially old versions of .xls files). Furthermore, the execution body of step 208 may generate a corresponding feature evaluation result report using the above-mentioned report generation tool according to the feature evaluation result report calculation method and output method of each feature verification subtask specified in the target verification task flow.

[0175] It should be noted that in the target verification task process, for each feature verification subtask, it is specified how to calculate and statistically analyze the results output by the feature verification subtask to generate a corresponding feature evaluation result report, and the output method of the result report is also specified. Therefore, feature mining analysts can obtain more and richer information through the above feature evaluation result report, which facilitates feature mining.

[0176] It should be noted that the above-mentioned various optional implementation modes can be combined at will.

[0177] For example, the optional implementation of step 204 regarding the scheduling policy may be combined with the optional implementation of step 207 regarding leaving traces for designated tasks.

[0178] For example, step 204 may be executed first, and then step 207 may be executed, that is, the target user may first specify the scheduling strategy, and then specify the task trace method; in this way, after obtaining the feature verification task trace description information input by the target user in step 207, an interactive note file may be generated based on the feature verification task description information obtained in step 201, the feature verification task scheduling strategy information obtained in step 204, and the feature verification task trace description information obtained in step 207, and in the subsequent step 202, the above-generated interactive note file including the feature verification task description information, the feature verification task scheduling strategy information, and the feature verification task trace description information may be converted into a feature verification task statement sequence.

[0179] For another example, step 207 may be executed first, and then step 204 may be executed, that is, the target user may first specify the task trace method, and then specify the scheduling strategy; in this way, after obtaining the feature verification task scheduling strategy information input by the target user in step 204, an interactive note file may be generated based on the feature verification task description information obtained in step 201, the feature verification task trace description information obtained in step 207, and the feature verification task scheduling strategy information obtained in step 204, and in the subsequent step 202, the above-generated interactive note file including the feature verification task description information, the feature verification task trace description information, and the feature verification task scheduling strategy information may be converted into a feature verification task statement sequence.

[0180] The feature verification task processing method provided by the above embodiment of the present disclosure first obtains the feature verification task description information input by the target user, wherein the feature verification task description information includes the training sample set description information, the target model description information, the target model parameter description information, the feature description information to be verified and the target verification task process description information, and the target verification task process is used to indicate the feature verification subtasks included in the feature verification task and the order between the feature verification subtasks, and the feature verification subtask is used to determine at least one feature evaluation index; then, a feature verification task statement sequence for executing the feature verification task is generated according to the feature verification task description information; finally, the feature verification task is executed based on the feature verification task statement sequence, where the feature verification task is used to execute each feature verification subtask according to the target verification task process, and in the process of executing each feature verification subtask, the training sample set is used as the sample data, and the target model corresponding to the target model parameter is used as the machine learning model to verify the feature evaluation index of the feature to be verified in each feature verification subtask. Thus, it is possible to realize the feature evaluation process of customizing the training sample set, the machine learning model, the model parameters and at least one feature evaluation index without manually writing code, simplify the evaluation process, save evaluation time, and facilitate management and control.

[0181] To achieve Figure 2 The feature verification task processing method shown in various optional implementations thereof can encapsulate each component of the feature verification task processing method into a sub-library file, and encapsulate the feature verification task processing method into a feature evaluation library file. Figure 2 For various optional implementations, please refer to Figure 3 , Figure 3 The main structure between each sub-library (or functional module) in the feature evaluation library file is shown.

[0182] like Figure 3 As shown, the library file architecture includes an input module based on Jupyter, a code conversion module based on nbconvert, a subtask execution module based on Jypter, a scheduling module based on celery, a log and monitoring module based on logging, and a report generation module based on openpyxl and / or xlrd. Each of the above modules can form a sub-library file with corresponding functions.

[0183] Among them, the input module is used to receive feature verification task description information, feature verification task trace description information, feature verification task scheduling strategy information and feature verification task log recording configuration information input by the target user using Jupyter Notebook.

[0184] After receiving the feature verification task description information or the feature verification task trace description information, the input module will send it to the code conversion module. After receiving the feature verification task scheduling strategy information, it will send it to the scheduling module. After receiving the feature verification task log record configuration information, it will send it to the log and monitoring module.

[0185] After receiving the feature verification task scheduling strategy information, the scheduling module will create a scheduling task and send the feature verification task scheduling strategy information to the code conversion module.

[0186] After receiving the logging configuration information of the feature verification task, the log and monitoring module will distribute the logging configuration of each feature verification subtask to the corresponding feature verification subtask.

[0187] After receiving the feature verification task description information, or after receiving the feature verification task trace description information and / or the feature verification task scheduling strategy information again based on the received feature verification task description information, the feature verification task trace description information and the feature verification task scheduling strategy information, the code conversion module generates a feature verification task statement sequence based on the received feature verification task description information, the feature verification task trace description information and the feature verification task scheduling strategy information, and sends the feature verification task statement sequence to the subtask execution module. Among them, the feature verification task description information is used by the code conversion module to select a training sample set from a data set, select a target model from a model library, select features to be verified from a feature library, select target model parameters from a parameter set, and select a target verification task process from a verification task process template.

[0188] The subtask execution module can execute the feature verification task according to the received feature verification task statement sequence. At the same time, if a log record configuration is received, a log file is generated for storage or output according to the corresponding log record configuration during the execution of each feature verification subtask.

[0189] The report generation module can receive the output results of each feature verification subtask output by the subtask execution module and generate a corresponding feature evaluation result report.

[0190] Further references Figure 4 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a feature verification task processing device, and the device embodiment is Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0191] like Figure 4As shown, the feature verification task processing device 400 of this embodiment includes: a first acquisition module 401, a generation module 402 and an execution module 403. Among them, the first acquisition module 401 is configured to obtain feature verification task description information input by the target user, the feature verification task description information includes training sample set description information, target model description information, target model parameter description information, to-be-verified feature description information and target verification task process description information, the target verification task process is used to indicate the feature verification subtasks included in the feature verification task and the sequence between the feature verification subtasks, wherein the feature verification subtask is used to determine at least one feature evaluation index; the generation module 402 is configured to generate a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, wherein the feature verification task statement sequence is a statement sequence based on a logical language; the execution module 403 is configured to execute the feature verification task based on the feature verification task statement sequence, wherein the feature verification task is used to execute each feature verification subtask according to the target verification task process, and in the process of executing each feature verification subtask, the training sample set is used as sample data, and the target model corresponding to the target model parameter is used as a machine learning model to verify the feature evaluation index of the to-be-verified feature in each of the feature verification subtasks.

[0192] In this embodiment, the specific processing of the first acquisition module 401, the generation module 402 and the execution module 403 of the feature verification task processing device 400 and the technical effects thereof can be referred to respectively. Figure 2 The relevant descriptions of step 201, step 202 and step 203 in the corresponding embodiment are not repeated here.

[0193] In some optional implementations, the device 400 may further include:

[0194] A second acquisition module 404 is configured to acquire, before generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, feature verification task scheduling strategy information input by the target user, wherein the feature verification task scheduling strategy information is used to indicate the execution time and / or execution order of feature verification subtasks in the feature verification task; and

[0195] The generating module 402 and the executing module 403 are further configured to:

[0196] A scheduling task is created and executed based on the feature verification task scheduling policy information, wherein the scheduling task is used to generate a feature verification subtask statement sequence for executing corresponding feature verification subtasks based on the feature verification task description information and the feature verification task scheduling policy information according to the execution time and / or execution order of each feature verification subtask indicated by the feature verification task scheduling policy information, and execute the corresponding feature verification subtask based on the feature verification subtask statement sequence.

[0197] In some optional implementations, the device 400 may further include:

[0198] A third acquisition module 405 is configured to acquire the feature verification task log recording configuration information input by the target user before generating the feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, wherein the feature verification task log recording configuration information is used to indicate the log recording configuration during the execution of the feature verification subtask in the feature verification task; and

[0199] The configuration distribution module 406 is configured to distribute the logging configuration of each feature verification subtask to the corresponding feature verification subtask before executing the feature verification task based on the feature verification task statement sequence.

[0200] In some optional implementations, the feature verification task is also used to output a log record of the execution process of each feature verification subtask.

[0201] In some optional implementations, the device 400 may further include:

[0202] A fourth acquisition module 407 is configured to obtain the feature verification task trace description information input by the target user before generating the feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, wherein the feature verification task trace description information is used to indicate the code position and trace method that need to be recorded during the execution of the feature verification subtask in the feature verification task; and

[0203] The generating module 402 may be further configured to generate a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information and the feature verification task trace description information.

[0204] In some optional implementations, the device 400 may further include:

[0205] The result generation module 408 is configured to generate a feature evaluation result report for the output result of each feature verification subtask after executing the feature verification task based on the feature verification task statement sequence.

[0206] It should be noted that the implementation details and technical effects of each module in the feature verification task processing device provided in the embodiments of the present disclosure can be referred to the description of other embodiments in the present disclosure and will not be repeated here.

[0207] Reference below Figure 5 , which shows a schematic diagram of the structure of a computer system 500 suitable for implementing the electronic device of the present disclosure. Figure 5 The computer system 500 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0208] like Figure 5 As shown, the computer system 500 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage device 508 to a random access memory (RAM) 503. Various programs and data required for the operation of the computer system 500 are also stored in the RAM 503. The processing device 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0209] Typically, the following devices may be connected to the I / O interface 505: input devices 506 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, etc.; output devices 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 508 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 509. The communication devices 509 may allow the computer system 500 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 5 The computer system 500 of an electronic device having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0210] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device 509, or installed from a storage device 508, or installed from a ROM 502. When the computer program is executed by the processing device 501, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0211] It should be noted that the computer-readable medium disclosed above 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, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0212] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0213] The computer readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device implements the following Figure 2The illustrated embodiment and its optional implementations illustrate a method for processing a feature verification task.

[0214] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages, such as Python, Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0215] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0216] The modules involved in the embodiments described in the present disclosure may be implemented by software or hardware. The name of a module does not limit the module itself in some cases. For example, the first acquisition module may also be described as a "module for acquiring feature verification task description information input by the target user."

[0217] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

Claims

1. A method for processing a feature verification task, comprising: Acquire feature verification task description information input by a target user, wherein the feature verification task description information includes training sample set description information, target model description information, target model parameter description information, feature description information to be verified, and target verification task process description information, wherein the target verification task process is used to indicate each feature verification subtask included in the feature verification task and the sequence between each feature verification subtask, wherein the feature verification subtask is used to determine at least one feature evaluation indicator; Generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, wherein the feature verification task statement sequence is a statement sequence based on a logical language; The feature verification task is performed based on the feature verification task statement sequence, wherein the feature verification task is used to execute each feature verification subtask according to the target verification task process, and in the process of executing each feature verification subtask, the training sample set is used as sample data, and the target model corresponding to the target model parameters is used as the machine learning model to verify the feature evaluation index of the feature to be verified in each of the feature verification subtasks.

2. The method according to claim 1, characterized in that Before generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, the method further includes: Acquire feature verification task scheduling strategy information input by the target user, wherein the feature verification task scheduling strategy information is used to indicate the execution time and / or execution order of feature verification subtasks in the feature verification task; and The step of generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, and executing the feature verification task based on the feature verification task statement sequence, comprises: A scheduling task is created and executed based on the feature verification task scheduling policy information, wherein the scheduling task is used to generate a feature verification subtask statement sequence for executing corresponding feature verification subtasks based on the feature verification task description information and the feature verification task scheduling policy information according to the execution time and / or execution order of each feature verification subtask indicated by the feature verification task scheduling policy information, and execute the corresponding feature verification subtask based on the feature verification subtask statement sequence.

3. The method according to claim 1, characterized in that: Before generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, the method further includes: Acquire feature verification task log record configuration information input by the target user, wherein the feature verification task log record configuration information is used to indicate the log record configuration during the execution of a feature verification subtask in the feature verification task; and Before executing the feature verification task based on the feature verification task statement sequence, the method further includes: Distribute the logging configuration of each feature verification subtask to the corresponding feature verification subtask.

4. The method according to claim 3, characterized in that The feature verification task is also used to output a log record of the execution process of each feature verification subtask.

5. The method according to claim 1, characterized in that: Before generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, the method further includes: Acquire the feature verification task trace description information input by the target user, wherein the feature verification task trace description information is used to indicate the code location and trace method that need to be recorded during the execution of the feature verification subtask in the feature verification task; and The step of generating a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information includes: A feature verification task statement sequence for executing the feature verification task is generated based on the feature verification task description information and the feature verification task trace description information.

6. The method according to claim 1, characterized in that After executing the feature verification task based on the feature verification task statement sequence, the method further includes: For the output results of each feature verification subtask, a feature evaluation result report is generated.

7. A feature verification task processing device, comprising: A first acquisition module is configured to acquire feature verification task description information input by a target user, wherein the feature verification task description information includes training sample set description information, target model description information, target model parameter description information, to-be-verified feature description information, and target verification task process description information, wherein the target verification task process is used to indicate each feature verification subtask included in the feature verification task and the sequence between each feature verification subtask, wherein the feature verification subtask is used to determine at least one feature evaluation indicator; A generating module, configured to generate a feature verification task statement sequence for executing the feature verification task based on the feature verification task description information, wherein the feature verification task statement sequence is a statement sequence based on a logical language; An execution module is configured to execute the feature verification task based on the feature verification task statement sequence, wherein the feature verification task is used to execute each feature verification subtask according to the target verification task process, and in the process of executing each feature verification subtask, the training sample set is used as sample data, and the target model corresponding to the target model parameters is used as a machine learning model to verify the feature evaluation index of the feature to be verified in each of the feature verification subtasks.

8. An electronic device comprising: one or more processors; a storage device having one or more programs stored thereon, When the one or more programs are executed by the one or more processors, the one or more processors are enabled to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, wherein: When the computer program is executed by one or more processors, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program / instruction, wherein when the computer program / instruction is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images