Feature management method and apparatus, computer device, storage medium, and product

Abstract

Embodiments of the present application disclose a feature management method and device, computer equipment, storage medium and product. Embodiments of the present application can obtain feature original data and description information of the feature original data, wherein the description information includes feature attribute information and feature evaluation information. According to the description information, a feature entry operation is performed on the feature original data. Based on the feature attribute information, the features after entry are classified to determine the feature categories to which the features after entry belong. Based on the feature evaluation information, feature evaluation indexes of the features after entry in at least one evaluation dimension are calculated. According to the feature categories and the feature evaluation indexes, the features after entry are managed. The scheme is applied to various scenes such as cloud technology, artificial intelligence and intelligent transportation, and can improve feature management efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a feature management method and device, computer equipment, storage medium and product. BACKGROUND

[0002] A feature is a special symbol or sign that can be used to identify a person or thing, for example, in the field of computers, features can be used to distinguish between people or things.

[0003] In the research and practice process of the related technology, the inventors of the present application found that the current feature management is relatively inefficient, for example, the applicable business scenarios are relatively limited, and the degree of coincidence with the whole process from production to application in actual application is not high, so that the efficiency of feature management needs to be improved. SUMMARY

[0004] The embodiments of the present application provide a feature management method and device, computer equipment, storage medium and product, which can improve the efficiency of feature management.

[0005] The embodiments of the present application provide a feature management method, comprising:

[0006] Obtaining feature raw data and description information of the feature raw data, wherein the description information comprises feature attribute information and feature evaluation information;

[0007] According to the description information, performing feature entry operation on the feature raw data;

[0008] Based on the feature attribute information, classifying the entered feature to determine the feature category to which the entered feature belongs;

[0009] Based on the feature evaluation information, calculating a feature evaluation index of the entered feature in at least one evaluation dimension;

[0010] According to the feature category and the feature evaluation index, performing feature management on the entered feature.

[0011] Correspondingly, the embodiments of the present application also provide a feature management device, comprising:

[0012] An obtaining unit is configured to obtain feature raw data and description information of the feature raw data, wherein the description information comprises feature attribute information and feature evaluation information;

[0013] An entry unit is configured to perform feature entry operation on the feature raw data according to the description information;

[0014] a classification unit, configured to perform feature classification on the post-entry feature based on the feature attribute information, to determine a feature category to which the post-entry feature belongs;

[0015] a calculation unit, configured to calculate a feature evaluation index of the post-entry feature in at least one evaluation dimension based on the feature evaluation information;

[0016] a first management unit, configured to perform feature management on the post-entry feature according to the feature category and the feature evaluation index.

[0017] In an embodiment, the feature evaluation information includes a feature evaluation sample set and at least one feature evaluation operator, each of the feature evaluation operators corresponding to one of the evaluation dimensions; and the calculation unit includes:

[0018] a result calculation sub-unit, configured to calculate a feature evaluation result of the post-entry feature in a corresponding evaluation dimension in the feature evaluation sample set according to the at least one feature evaluation operator;

[0019] a result conversion sub-unit, configured to convert the feature evaluation result into the feature evaluation index of the post-entry feature in at least one evaluation dimension according to a preset feature evaluation threshold.

[0020] In an embodiment, the entry unit includes:

[0021] an information extraction sub-unit, configured to extract meta-information of the feature original data from the description information;

[0022] a relationship establishment sub-unit, configured to establish an association relationship between the meta-information and the feature original data;

[0023] a feature entry sub-unit, configured to perform a feature entry operation on the feature original data based on the association relationship.

[0024] In an embodiment, the first management unit includes:

[0025] a feature grouping sub-unit, configured to perform feature grouping on the post-entry feature based on the feature category;

[0026] a combination determination sub-unit, configured to determine a feature combination to which the post-entry feature belongs according to a grouping result, wherein the feature combination includes at least one post-entry feature belonging to the same feature category;

[0027] a first management sub-unit, configured to perform feature management on the post-entry feature according to the feature combination and the feature evaluation index.

[0028] In an embodiment, the feature attribute information comprises feature classification information; and the classification unit comprises:

[0029] a candidate determination subunit configured to determine, based on the feature classification information, a candidate feature category matching the post-entry feature from a preset category set;

[0030] a matching calculation subunit configured to calculate a matching degree between the post-entry feature and the candidate feature category;

[0031] a category determination subunit configured to determine, according to the matching degree, a feature category to which the post-entry feature belongs.

[0032] In an embodiment, the description information further comprises feature derivation information, the feature derivation information representing that the post-entry feature is generated based on feature derivation, and the feature management apparatus further comprises:

[0033] a native determination unit configured to determine, according to the feature derivation information, a native feature having a derivation relationship with the post-entry feature, wherein the post-entry feature is generated based on the native feature;

[0034] a second management unit configured to perform feature management on the post-entry feature based on the derivation relationship and the native feature.

[0035] In an embodiment, the second management unit comprises:

[0036] a page display subunit configured to display a feature derivation schematic page of the post-entry feature based on the derivation relationship, wherein the feature derivation schematic page comprises a feature derivation schematic diagram, and the feature derivation schematic diagram comprises a feature node corresponding to the native feature and a feature node corresponding to the post-entry feature.

[0037] an information display subunit configured to display meta information corresponding to a target post-entry feature on the feature derivation schematic page in response to a node selection operation on a target feature node, so as to perform feature management on the target post-entry feature, wherein the target post-entry feature is a post-entry feature corresponding to the target node.

[0038] In an embodiment, the description information further comprises feature permission information, the feature permission information representing feature access permissions of a user group to the post-entry feature, and the feature management apparatus further comprises:

[0039] a group determination unit configured to determine, according to the feature permission information, a target user group corresponding to the post-entry feature, wherein the target user group comprises at least one user, and each of the users has feature access permissions to the post-entry feature;

[0040] A third management unit is configured to manage the post-entry features by the target user group.

[0041] In an embodiment, the third management unit comprises:

[0042] A set determination sub-unit is configured to determine a set of accessible features of the target user group, wherein the set of accessible features comprises at least one accessible post-entry feature of the target user group, and a user in the target user group has feature access right to each of the accessible post-entry features.

[0043] A feature determination sub-unit is configured to determine a target post-entry feature to be adjusted from the set of accessible features.

[0044] A feature adjustment sub-unit is configured to perform feature adjustment on the set of accessible features based on the target post-entry feature, so as to adjust the feature access right of the target user group.

[0045] A second management sub-unit is configured to manage the post-entry features by the target user group after adjustment.

[0046] In an embodiment, the third management unit comprises:

[0047] A user determination sub-unit is configured to determine a target user to be adjusted from the target user group.

[0048] A user adjustment sub-unit is configured to perform user adjustment on the target user group based on the target user, so as to adjust the target user group.

[0049] A third management sub-unit is configured to manage the post-entry features by the target user group after adjustment.

[0050] Correspondingly, the embodiments of the present application further provide a storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the steps of the feature management method shown in the embodiments of the present application.

[0051] Correspondingly, the embodiments of the present application further provide a computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the feature management method shown in the embodiments of the present application.

[0052] Correspondingly, the embodiments of the present application further provide a computer program product comprising computer programs / instructions, wherein the computer programs / instructions are executed by a processor to implement the steps of the feature management method shown in the embodiments of the present application.

[0053] The embodiment of the present application can obtain feature raw data and description information of the feature raw data, wherein the description information comprises feature attribute information and feature evaluation information; perform feature entry operation on the feature raw data according to the description information; perform feature classification on the entered feature based on the feature attribute information to determine a feature category to which the entered feature belongs; calculate a feature evaluation index of the entered feature in at least one evaluation dimension based on the feature evaluation information; and perform feature management on the entered feature according to the feature category and the feature evaluation index.

[0054] The scheme considers multiple links from production to application of features in actual application when managing the features, for example, the scheme comprises links such as obtaining feature raw data, feature entry, feature evaluation, and feature management, which can greatly improve the efficiency of managing the features. Specifically, the scheme can perform feature entry operation according to the description information of the feature raw data, which can orderly enter the features, and facilitate subsequent feature classification, feature query, feature explanation, and the like based on the description information; and the scheme can calculate a feature evaluation index of the entered feature in at least one evaluation dimension, so that the quality of the entered feature can be controlled from multiple aspects, thereby reducing accidents caused by feature quality problems in the feature management process and improving the robustness of the feature management system. In addition, the scheme can be applied to a wide range of business scenarios, for example, the scheme can be applied to feature management of user portrait features, financial modeling features, and features in a recommendation scenario, so that the scheme can be applied to more extensive business scenarios in actual application and can improve the feature management efficiency from a larger coverage. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0056] Figure 1 is a scene diagram of the feature management method provided by the embodiment of the present application;

[0057] Figure 2 is a flowchart of the feature management method provided by the embodiment of the present application;

[0058] Figure 3 is a feature entry diagram of the feature management method provided by the embodiment of the present application;

[0059] Figure 4is a feature grouping schematic diagram of a feature management method provided by an embodiment of the application;

[0060] Figure 5 is a feature quality monitoring schematic diagram of a feature management method provided by an embodiment of the application;

[0061] Figure 6 is a feature derivation schematic diagram of a feature management method provided by an embodiment of the application;

[0062] Figure 7 is a user group management schematic diagram of a feature management method provided by an embodiment of the application;

[0063] Figure 8 is a feature adjustment schematic diagram of a feature management method provided by an embodiment of the application;

[0064] Figure 9 is a user adjustment schematic diagram of a feature management method provided by an embodiment of the application;

[0065] Figure 10 is another flow schematic diagram of a feature management method provided by an embodiment of the application;

[0066] Figure 11 is an architecture design schematic diagram of a feature management method provided by an embodiment of the application;

[0067] Figure 12 is a feature search schematic diagram of a feature management method provided by an embodiment of the application;

[0068] Figure 13 is a technical solution flow schematic diagram of a feature management method provided by an embodiment of the application;

[0069] Figure 14 is a structure schematic diagram of a feature management device provided by an embodiment of the application;

[0070] Figure 15 is another structure schematic diagram of a feature management device provided by an embodiment of the application;

[0071] Figure 16 is another structure schematic diagram of a feature management device provided by an embodiment of the application;

[0072] Figure 17 is another structure schematic diagram of a feature management device provided by an embodiment of the application;

[0073] Figure 18 is another structure schematic diagram of a feature management device provided by an embodiment of the application;

[0074] Figure 19 is another structure schematic diagram of a feature management device provided by an embodiment of the application;

[0075] Figure 20 is another structural schematic diagram of the feature management apparatus provided by an embodiment of the present application.

[0076] Figure 21 is another structural schematic diagram of the feature management apparatus provided by an embodiment of the present application.

[0077] Figure 22 is another structural schematic diagram of the feature management apparatus provided by an embodiment of the present application.

[0078] Figure 23 is another structural schematic diagram of the feature management apparatus provided by an embodiment of the present application.

[0079] Figure 24 is a structural schematic diagram of a computer device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0080] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0081] The embodiments of the present application provide a feature management method, apparatus, computer device, storage medium and product. Specifically, the embodiments of the present application provide a feature management apparatus suitable for a computer device, where the computer device can be a server or a terminal, etc. Specifically, the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers; the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a vehicle-mounted computer, a vehicle-mounted computer, etc., but is not limited thereto. The terminal and the server can be directly or indirectly connected through wired or wireless communication, which is not limited in the present application.

[0082] The embodiments of the present application will take the feature management apparatus as an example of a terminal to introduce the feature management method.

[0083] Reference Figure 1The terminal 10 can obtain the feature raw data and the description information of the feature raw data, where the description information includes the feature attribute information and the feature evaluation information. For example, the terminal 10 can obtain the feature raw data and the description information of the feature raw data by interacting with the server 20, such as making a data request. Further, the terminal 10 can perform a feature entry operation on the feature raw data according to the description information, perform feature classification on the entered feature based on the feature attribute information to determine the feature category to which the entered feature belongs, and calculate the feature evaluation index of the entered feature in at least one evaluation dimension based on the feature evaluation information. In this way, the terminal 10 can perform feature management on the entered feature according to the feature category and the feature evaluation index, for example, the terminal 10 can upload the entered feature to the server 20, for another example, the terminal 10 can receive an instruction sent by the server 20 and perform feature management on the entered feature based on the instruction, and the like.

[0084] The following will be described in detail respectively. It should be noted that the order of the following embodiments is not limited as the preferred order of the embodiments.

[0085] The feature management method provided by the embodiments of the present application can be executed by a server or a terminal, or can be executed by a server and a terminal together. The embodiments of the present application take the case that the feature management method is executed by a terminal as an example to be described, for example, Figure 2 The specific process of the feature management method can be as follows:

[0086] 101. Obtain feature raw data and description information of the feature raw data, where the description information includes feature attribute information and feature evaluation information.

[0087] The feature is a special symbol or sign that can be identified by a person or thing, and is mainly used to distinguish between people or things in the computer field.

[0088] For example, in the application scenario of constructing a user portrait, the feature can include attribute features of the user, and the attributes of the user can include, for example, gender, age, city, etc. The feature can include behavior features of the user, and the behavior of the user can include, for example, purchase behavior, browsing behavior, collection behavior, and like behavior.

[0089] For another example, in the application scenario of financial modeling, the feature can include attribute features of a financial object, and the financial object can include, for example, a financial institution, a user, a company organization, etc. The attribute feature can specifically include a feature composed of basic attribute information of the financial object. The feature can include behavior features of the financial object, and the behavior feature can specifically include a feature composed of financial behavior information of the financial object.

[0090] For example, in the application scenario of content recommendation, the feature can include the attribute feature of the payment object, which can include a merchant organization and a payment user, and the attribute feature can include the feature composed of the basic attribute information of the payment object.

[0091] For example, the feature raw data can include the raw data of the attribute feature, the raw data of the behavior feature, and the like. For example, in the application scenario of financial modeling, when the feature is the behavior feature of a financial object, such as a transaction feature, the feature raw data corresponding to the transaction feature can include transaction financial. For another example, in the application scenario of constructing a user portrait, when the feature is the attribute feature of a user, such as an age feature, the feature raw data corresponding to the age feature can include the age data of the user.

[0092] For example, the description information of the feature raw data can include the related information for describing the feature corresponding to the feature raw data, such as the basic attribute of the feature, the role of the feature, how the feature is produced, and the like.

[0093] For example, the description information of the feature raw data can include the related information for describing the feature corresponding to the feature raw data, such as the basic attribute of the feature, the role of the feature, how the feature is produced, and the like. Figure 3 For example, the description information of the feature raw data can include the related information for describing the feature corresponding to the feature raw data, such as the basic attribute of the feature, the role of the feature, how the feature is produced, and the like.

[0094] For example, the description information of the feature raw data can include the related information for describing the feature corresponding to the feature raw data, such as the basic attribute of the feature, the role of the feature, how the feature is produced, and the like. Figure 3The description information listed above is only an example, and in actual applications, the description information can include more. For example, the description information can specifically include related information such as feature classification information, feature grouping information, feature derivation type, feature evaluation set, feature evaluation set magnitude, and feature evaluation set description. Among them, the "feature classification information" can be used to describe the basis for classifying the feature, for example, the feature classification information can specifically be a feature topic; the "feature grouping information" can be used to describe the basis for grouping the feature, for example, the feature grouping information can specifically be a feature group name; the "feature derivation type" can be used to describe whether the feature is derived from other features and the type of derivation method, for example, the feature derivation type can specifically include a native feature (i.e., a brand new feature that is not derived from other features), a financial derivation (i.e., a feature derived from a financial feature), a portrait derivation (i.e., a feature derived from a portrait feature), and the like; the "feature evaluation set" can be a data set used for evaluating the feature, for example, the feature evaluation set can specifically be a number of data records including the feature; the "feature evaluation set magnitude" can be a magnitude used to describe the evaluation set of the feature; and the "feature evaluation set description" can be used as the description information of the evaluation set of the feature.

[0095] Among them, the feature attribute information is related information used to describe the attribute of the feature corresponding to the feature raw data, for example, the feature attribute information can include a feature name, a feature type, a feature description, feature classification information, a feature group name, and the like.

[0096] Among them, the feature evaluation information is related information required for evaluating the feature corresponding to the feature raw data, for example, the feature evaluation information can include quality monitoring, a feature evaluation set, a feature evaluation set magnitude, a feature evaluation set description, and the like.

[0097] There can be various ways for the terminal to obtain the feature raw data and the description information of the feature raw data, for example, the server can send the feature raw data and the description information of the feature raw data to the terminal to enable the terminal to obtain the feature raw data and the description information of the feature raw data; for another example, the terminal can obtain the feature raw data and the description information of the feature raw data by querying the data stored locally; for another example, the terminal can obtain the feature raw data and the description information of the feature raw data by requesting other terminals; for another example, the terminal can obtain the feature raw data and the description information of the feature raw data based on the interaction between the terminal and the user; and the like.

[0098] In an embodiment, the terminal can display a feature acquisition page, where the feature acquisition page can include a feature selection region and a data acquisition region, the feature selection region can include at least one feature selection control, each feature selection control corresponds to a to-be-entered feature, and the data acquisition region is used for the terminal to acquire feature raw data corresponding to the to-be-entered feature and description information of the feature raw data; in response to a selection operation on a target feature selection control, at least one data acquisition control is displayed in the data acquisition region, the data acquisition control is used to acquire target data of the target to-be-entered feature, and the target data includes the feature raw data of the to-be-entered feature and the description information of the feature raw data; based on a trigger operation on the target data acquisition control, the feature raw data and the description information of the feature raw data are acquired.

[0099] The implementation of the step "based on the trigger operation on the target data acquisition control, acquiring the feature raw data and the description information of the feature raw data" can be various, for example, the step can include: based on the trigger operation on the target data acquisition control, acquiring the feature raw data and the description information of the feature raw data; or for example, the feature acquisition page can further include a data submission control, and the step can include: in response to the trigger operation on the target data acquisition control, determining to-be-submitted data, where the to-be-submitted data includes the feature raw data and the description information of the feature raw data; and in response to a data submission operation on the data submission control, submitting the to-be-submitted data to the terminal, so that the terminal acquires the feature raw data and the description information of the feature raw data.

[0100] It is worth noting that the forms of the interactive operations involved in the embodiments of the present application can be specific touch operations, such as single-click operation, long-press operation, double-click operation, and sliding operation, etc., and can also be triggered by voice; optionally, the trigger operation can also be a combination of a series of operations, and the embodiments of the present application do not limit this, and the present application will not be described again.

[0101] As an example, referring to Figure 3 , the terminal can display the feature acquisition page shown in 1001, where the feature acquisition page can include the feature selection region shown in 1002 and the data acquisition region shown in 1003, the feature selection region 1002 can include at least one feature selection control, and in Figure 3 , the feature selection control is a control including the "field" string, each feature selection control corresponds to a to-be-entered feature, and the data acquisition region 1003 is used for the terminal to acquire feature raw data corresponding to the to-be-entered feature and description information of the feature raw data.

[0102] Further, in this example, the user can determine the target feature selection control in the feature selection region 1002 by a selection operation, for example, by selecting the target feature selection control 1004. Figure 3 In this example, the user can select the target feature selection control 1004 by a click. In response to the selection operation on the target feature selection control 1004, the terminal can display at least one data acquisition control in the data acquisition region 1003, for example, including controls for acquiring data of feature name, feature type, feature description, quality monitoring, feature source, production caliber, visible range, etc., which are used for the user to upload target data of the target feature to be entered, including feature raw data of the feature to be entered and description information of the feature raw data.

[0103] In this example, the user can implement the triggering operation on the target data acquisition control by inputting in the control or selecting in the list, and the terminal can correspondingly acquire the feature raw data and the description information of the feature raw data in response to the triggering operation. In actual application, the terminal can acquire the feature raw data in multiple ways in response to the triggering operation, for example, the terminal can acquire the data corresponding after the user provides the relevant data to the terminal by the triggering operation; for another example, the feature acquisition page 1001 can include a data submission control 1005, the user can submit the relevant data to the terminal by triggering the data submission control 1005, and correspondingly, the terminal can acquire the relevant data, i.e., the feature raw data and the description information of the feature raw data, in response to the data submission operation on the data submission control 1005.

[0104] 102. Perform a feature entry operation on the feature raw data according to the description information.

[0105] The feature entry operation refers to an operation of entering a feature in a feature management system. In this application, since the feature raw data is the raw data of the feature, and the description information of the feature raw data is the relevant information for describing the feature, the feature entry operation can be performed on the feature raw data according to the description information of the feature raw data to enter the corresponding feature in the feature management system.

[0106] In an embodiment, the feature entry operation can be implemented based on the interaction with the user. For example, the terminal can perform the feature entry operation on the feature raw data in response to the data submission operation on the data submission control. As an example, the user can upload the feature raw data and the description information of the feature raw data by triggering the data acquisition control. Further, the user can trigger the terminal to perform the feature entry operation on the feature raw data according to the description information by triggering the data submission control, and correspondingly, the terminal can perform the feature entry operation on the feature raw data according to the description information in response to the data submission operation on the data submission control.

[0107] In another embodiment, considering that the description information can be used to describe the feature corresponding to the feature raw data, therefore, when the feature is entered, the correlation between the description information and the feature raw data can be considered to make the feature entry operation more accurate and reasonable. Specifically, the step of "performing feature entry operation on the feature raw data according to the description information" can include:

[0108] extracting meta information of the feature raw data from the description information;

[0109] establishing an association relationship between the meta information and the feature raw data;

[0110] performing feature entry operation on the feature raw data based on the association relationship.

[0111] Wherein, the meta information is information about information, which can be used to describe the structure, semantics, purpose and usage of information, etc. Specifically, since the description information of the feature raw data is used to describe the feature corresponding to the feature raw data, in actual application, the description information can be related to the original information uploaded by the user, therefore, the terminal can further extract the meta information of the feature raw data from the description information on demand.

[0112] As an example, the description information of the feature raw data can include information describing feature name, feature type, feature description, quality monitoring, feature source, production caliber, visible range, feature classification information, feature grouping information, feature derivation type, feature evaluation set, feature evaluation set magnitude, and feature evaluation set description, etc. The terminal can extract part or all of the information from the description information on demand as the meta information of the feature raw data.

[0113] Further, an association relationship between the meta information and the feature raw data can be established. There can be various ways to do this. In an embodiment, a mapping relationship between the meta information and the feature raw data can be established to establish an association relationship between them, for example, a mapping table can be created in a database to record the mapping relationship between them, or a data structure for implementing mapping can be applied to establish the mapping relationship between them, etc.

[0114] In this way, the terminal can perform feature entry operation on the feature raw data based on the association between the feature raw data and the meta information of the feature raw data. For example, the feature management system can maintain a feature table, which can be used to record the feature raw data of each feature. The terminal can add the feature raw data to the feature table to implement feature entry of the feature raw data. It is worth noting that in actual application, the association between the meta information and the feature raw data is considered, and therefore, the meta information of the feature raw data and the feature raw data can be entered into the feature management system based on the association, so that when the entered feature is managed, the corresponding meta information of the entered feature can be queried and fully utilized to expand the function and efficiency of the feature management system. In addition, after the feature entry operation is performed on the feature raw data, the feature raw data can be registered as an entered feature available for users to use.

[0115] 103. performing feature classification on the entered feature based on the feature attribute information to determine a feature category to which the entered feature belongs.

[0116] In the present application, after the feature entry operation is performed on the feature raw data based on the description information, the feature raw data can be registered as an entered feature available for users to use.

[0117] In the present application, the feature classification refers to classifying the entered feature to determine the feature category to which the entered feature belongs.

[0118] In the present application, since the feature attribute information can be used to describe the attribute of the entered feature corresponding to the feature raw data, and the number of feature attribute information can be at least one, i.e., at least one feature attribute corresponds to relevant information, when the entered feature is classified based on the feature attribute information, a plurality of feature attribute information can be selected and used as the basis for classifying the entered feature.

[0119] As an example, the feature attribute information can include feature name, feature type, feature description, feature classification information, feature group name, etc. For example, a single feature attribute can be used as the basis for classifying the entered feature. For example, the entered feature can be classified according to the feature type, so that the entered features belonging to the same feature type belong to the same feature category.

[0120] For another example, at least two feature attributes can be used as the basis for classifying the entered feature. For example, different feature attributes can be assigned different priorities or weights, and the entered feature can be classified based on the priorities or weights of the feature attributes.

[0121] In an embodiment, the feature attribute information can specifically include feature classification information, wherein the feature classification information can be used to describe the basis for classifying the post-entry feature, and specifically, the step of "classifying the post-entry feature based on the feature attribute information to determine the feature category to which the post-entry feature belongs" can include:

[0122] determining a candidate feature category matching the post-entry feature from the preset category set based on the feature classification information;

[0123] calculating the matching degree between the post-entry feature and the candidate feature category;

[0124] determining the feature category to which the post-entry feature belongs according to the matching degree.

[0125] The preset category set is a set composed of at least one feature category. In the present application, each feature category can have at least one post-entry feature belonging to the feature category, and each post-entry feature can belong to multiple feature categories.

[0126] The manner of determining the candidate feature category matching the post-entry feature from the preset category set based on the feature classification information corresponding to the post-entry feature can be various, for example, the feature category related to the feature classification information in the preset category set can be taken as the candidate feature category. The manner of calculating the correlation between the feature category and the feature classification information can be various, for example, the semantic similarity between the two can be calculated to determine the correlation between the two; for example, each feature category in the preset feature set can have at least one label, and correspondingly, the feature classification information can be assigned a label, and the correlation between the two can be determined by calculating the coincidence degree of the labels.

[0127] Further, the matching degree between the post-entry feature and the candidate feature category can be calculated to enable the determination of the feature category to which the post-entry feature belongs from the candidate feature category based on the calculated matching degree. The manner of calculating the matching degree between the post-entry feature and the candidate feature category can be various, for example, each feature category in the preset feature set can have at least one label, the matching degree between the post-entry feature and each label of the feature category (for distinction, this matching degree can be referred to as the first matching degree) can be calculated, and the first matching degree can be aggregated to obtain the matching degree between the post-entry feature and the feature category (for distinction, this matching degree can be referred to as the second matching degree), such as by statistical analysis of the first matching degree to obtain the second matching degree.

[0128] For example, a vector corresponding to the post-entry feature (for the sake of distinction, the vector can be referred to as a first vector) and a vector corresponding to the candidate feature category (for the sake of distinction, the vector can be referred to as a second vector) can be constructed respectively, and the matching degree between the post-entry feature and the candidate feature category can be determined by calculating the distance between the first vector and the second vector. For example, the first vector can be generated based on the meta information of the post-entry feature, and the second vector can be generated based on the name or label information of the candidate feature category.

[0129] According to the calculated matching degree, the feature category to which the post-entry feature belongs can be determined from the candidate feature categories in various ways. For example, the candidate feature category with the highest matching degree can be determined as the feature category to which the post-entry feature belongs. For another example, the candidate feature category with the highest matching degree within a preset matching degree interval range can be determined as the feature category to which the post-entry feature belongs. And so on.

[0130] 104. Based on the feature evaluation information, calculate the feature evaluation index of the post-entry feature in at least one evaluation dimension.

[0131] The feature evaluation information is relevant information required for evaluating the post-entry feature. For example, the feature evaluation information can include quality monitoring, feature evaluation set, feature evaluation set magnitude, feature evaluation set description, etc.

[0132] The evaluation dimension refers to the thinking dimension for evaluating the post-entry feature. Therefore, by calculating the feature evaluation index of the post-entry feature in at least one evaluation dimension, the post-entry feature can be evaluated from at least one evaluation dimension.

[0133] The feature evaluation index is a parameter for measuring the evaluation result of the post-entry feature. The feature evaluation index of the post-entry feature in at least one evaluation dimension respectively represents the feature evaluation result of the post-entry feature in the corresponding evaluation dimension.

[0134] In the present application, similar to the feature attribute information, the feature evaluation information corresponding to the post-entry feature can include a plurality of information required for feature evaluation on the post-entry feature, for example, the feature evaluation information can include quality monitoring information, and the quality monitoring information can specifically include at least one strategy information when performing quality monitoring or feature evaluation on the post-entry feature, for example, the strategy information can specifically be a feature evaluation operator. Therefore, a plurality of target strategy information can be selected from the at least one strategy information, and based on the target strategy information, a feature evaluation indicator of the post-entry feature can be calculated from at least one evaluation dimension. For example, each target strategy information can correspond to an evaluation dimension, and the feature evaluation indicator of the post-entry feature in the corresponding evaluation dimension can be calculated according to the calculation strategy of each target strategy information; for another example, one target strategy information or at least two strategy information can correspond to an evaluation dimension as needed, and the feature evaluation indicator of the post-entry feature in the corresponding evaluation dimension can be calculated according to the calculation strategy of the corresponding target strategy information; and the like.

[0135] In an embodiment, the feature evaluation information can include a feature evaluation sample set and at least one feature evaluation operator, and each feature evaluation operator corresponds to an evaluation dimension. Specifically, the step of "calculating a feature evaluation indicator of the post-entry feature in at least one evaluation dimension based on the feature evaluation information" can include:

[0136] According to the at least one feature evaluation operator, a feature evaluation result of the post-entry feature in the corresponding evaluation dimension is calculated in the feature evaluation sample set;

[0137] According to a preset feature evaluation threshold, the feature evaluation result is converted into a feature evaluation indicator of the post-entry feature in at least one evaluation dimension.

[0138] Wherein, the operator is a mapping from a function space to a function space, in practical application, the operator can be generalized to any space, and the operator can have input and output when used, and the operator then completes the conversion of the corresponding data.

[0139] In the present application, the feature evaluation operator is an operator for performing feature evaluation on the post-entry feature. For example, the feature evaluation operator can include: a population stability index (PSI), a zero value rate, a null value rate, a zero value fluctuation rate, a null value fluctuation rate, a partition integrity, and the like. Among them, the PSI is an important index for measuring the offset caused by the change of the feature sample, and is usually used to measure the stability of the feature sample; the zero value rate is used to measure the zero drop of the feature sample, the more the data zero drop, the greater the zero value rate, and the value range of the zero value rate is specifically [0, 1]; the null value rate is used to measure the abnormal situation that the feature sample should not be null but is null; the zero value fluctuation rate is used to detect the zero value rate fluctuation, for example, to view the data zero value rate trend; the null value fluctuation rate is used to detect the null value rate fluctuation, for example, to view the data null value rate trend; the partition integrity is used to test whether the feature partition of the partition table where the feature is located is complete, and is usually used to measure the data continuity and reliability.

[0140] Among them, the feature evaluation sample set is a data set used for feature evaluation on the post-entry feature, and the feature evaluation sample set can include at least one data record corresponding to a feature evaluation sample, wherein the data record can include a post-entry feature. For example, the feature evaluation sample can be specifically a sample user.

[0141] In the present application, the calculation method of the feature evaluation operator can be applied in the feature evaluation sample set, so as to calculate the feature evaluation result of the post-entry feature under the corresponding evaluation dimension. For example, the PSI, the zero value rate, the null value rate, the zero value fluctuation rate, the null value fluctuation rate, and the partition integrity of the post-entry feature can be calculated in the feature evaluation sample set respectively, and the calculation results are taken as the feature evaluation results of the post-entry feature under the corresponding evaluation dimension.

[0142] Among them, the preset feature evaluation threshold is a threshold for data conversion of the feature evaluation result of the post-entry feature. It is worth noting that different preset feature evaluation thresholds can be configured for the feature evaluation results under different evaluation dimensions. The form of the preset feature evaluation threshold can be various, for example, it can be a numerical value, and for example, it can be an interval range, and the like.

[0143] Since the feature evaluation result is the original result calculated according to the feature evaluation operator, in actual application, the feature evaluation result can be converted into a feature evaluation index through the preset feature evaluation threshold, so that the user can learn the evaluation performance of the post-entry feature under the corresponding evaluation dimension more clearly and directly according to the feature evaluation index.

[0144] The conversion manner can be various, for example, the preset feature evaluation threshold value can be assigned with the minimum weight (for example, 0), and the feature evaluation result of the post-entry feature can be taken as the feature evaluation index of the post-entry feature, so that the relevant user (for example, a development engineer) in the business scope can directly know the evaluation performance of the post-entry feature under the corresponding evaluation dimension according to the original feature evaluation result obtained by the feature evaluation operator.

[0145] For another example, the preset feature evaluation threshold value can be assigned with the maximum weight (for example, 1), the feature evaluation result can be compared with the preset feature evaluation threshold value, and the feature evaluation index of the post-entry feature can be determined according to the comparison result. For example, the comparison result can include whether the feature evaluation index meets the standard (for example, meets the standard or does not meet the standard), the corresponding grade of the feature evaluation index (for example, which can be presented in the form of a progress bar or a pie chart), and the like.

[0146] 105. According to the feature category and the feature evaluation index, the post-entry feature is subjected to feature management.

[0147] The feature management of the post-entry feature refers to the process of effectively organizing and controlling the post-entry feature. For example, the feature management of the post-entry feature can include grouping the post-entry feature, querying the feature derivation relationship between the post-entry features, querying the meta information of the post-entry feature, setting the relevant permissions of the post-entry feature, and the like.

[0148] In the present application, since the feature category to which the post-entry feature belongs can be determined by classifying the post-entry feature, and the feature evaluation index of the post-entry feature under at least one evaluation dimension can be calculated so that the user can monitor the quality of the post-entry feature from at least one evaluation dimension, the post-entry feature can be subjected to feature management according to the feature category and the feature evaluation index of the post-entry feature.

[0149] In an embodiment, the post-entry feature can be grouped according to the feature category of the post-entry feature, so that the user can quickly find the required feature from the feature management system. Specifically, the step of “according to the feature category and the feature evaluation index, the post-entry feature is subjected to feature management” can include:

[0150] The post-entry feature is subjected to feature grouping based on the feature category;

[0151] According to the grouping result, the feature combination to which the post-entry feature belongs is determined, wherein the feature combination includes at least one post-entry feature belonging to the same feature category;

[0152] According to the feature combination and the feature evaluation index, the post-entry feature is subjected to feature management.

[0153] The feature combination refers to a combination of at least one post-entry feature. In this application, the post-entry features can be grouped into feature combinations based on feature categories, so that post-entry features belonging to the same feature category can be grouped into the same feature combination.

[0154] It is worth noting that the correspondence between the post-entry features and the feature categories is not limited in this application, for example, each post-entry feature can only belong to one feature category, and each feature category can have multiple different post-entry features; for example, each post-entry feature can belong to multiple feature categories, and each feature category can have multiple different post-entry features; and so on.

[0155] In this application, the feature categories can be divided in various ways, for example, different themes can be determined according to the feature usage scenarios in actual applications, and different themes can be used as different feature categories. It is worth noting that the feature categories can have a hierarchical relationship, that is, a primary feature category can include several secondary categories, a secondary category can include several tertiary categories, and so on.

[0156] After determining the feature combination in which the post-entry feature is located, the post-entry feature can be managed according to the feature combination and the feature evaluation index.

[0157] As an example, the terminal can display a feature grouping page, wherein the feature grouping page includes a category hierarchy display area and a feature combination display area, the category hierarchy display area is used to display feature categories and the hierarchical relationship between the feature categories and other feature categories, and the feature combination display area is used to display feature combinations, each feature combination includes at least one post-entry feature in the same feature category; in response to a selection operation on a target feature category, at least one feature combination is displayed in the feature combination display area, wherein the feature combination includes at least one post-entry feature in the target feature category.

[0158] In this way, the terminal can display the feature grouping page to the user, so that the user can quickly find the feature combination under the target feature category, thereby greatly improving the efficiency of feature searching.

[0159] For example, different themes can be determined according to the feature usage scenarios in actual applications, and different themes can be used as different feature categories. Referring to Figure 4The terminal may display the feature grouping page shown in 1006, wherein the feature grouping page 1006 may include a category hierarchy display area 1007 and a feature combination display area 1008. The category hierarchy display area 1007 may display the hierarchical relationship between topics in the form of a tree diagram, and the feature combination display area 1008 may be used to display feature combinations. The user may select a target feature category from the category hierarchy display area. Correspondingly, the terminal may display at least one feature combination in the feature combination display area 1008 in response to the selection operation, wherein the feature combination includes at least one post-entry feature under the target feature category.

[0160] As another example, the terminal may display a feature quality monitoring page in response to a quality monitoring operation on a post-entry feature, wherein the feature quality monitoring page includes a feature quality monitoring list, the feature quality monitoring list includes feature quality monitoring data records corresponding to the post-entry feature, and each feature quality monitoring record includes a feature evaluation indicator of the post-entry feature under at least one evaluation dimension.

[0161] As an example, the terminal may display the following information in response to the quality monitoring operation for the post-entry features: Figure 5 The feature quality monitoring page 1009 shown, wherein the feature quality monitoring page 1009 may include a feature quality monitoring list 10010, and the feature quality monitoring list 10010 may include feature quality monitoring data records corresponding to the entered feature, and each feature quality monitoring record includes the feature evaluation index of the entered feature under at least one evaluation dimension, that is, the numerical values ​​corresponding to "monitoring operator 1", "monitoring operator 2", "monitoring operator 3", and "monitoring operator 4".

[0162] Optionally, in actual applications, the feature quality monitoring list may include more content than the above, for example, see Figure 5 The feature quality monitoring list can also display the feature name, library table name, field name, error cause, etc. of the entered feature, and provide more interactive operations, such as providing a "view" interactive operation for users to view the detailed quality monitoring information of the entered feature.

[0163] In practical applications, the description information may further include feature derivative information, which indicates that the feature is generated based on the feature derivative after being entered. Specifically, the feature management method may further include:

[0164] Determine, based on the feature derivative information, a native feature that has a derivative relationship with the recorded feature, wherein the recorded feature is derived based on the native feature;

[0165] Perform feature management on the entered features based on derived relationships and native features.

[0166] The feature derivation information is information used to describe that the post-entry feature is generated based on feature derivation. For example, the feature derivation information can be the "feature source" information described above.

[0167] The feature derivation refers to a process of generating a derived feature based on a native feature. For example, if post-entry feature 2 is generated based on post-entry feature 1, in this process, post-entry feature 1 is the native feature, and post-entry feature 2 is the derived feature derived from post-entry feature 1. In actual applications, the feature derivation process can be applied multiple times. For example, post-entry feature 2 can be generated based on post-entry feature 1, and then post-entry feature 3 can be generated based on post-entry feature 2, and so on, until post-entry feature N is generated based on post-entry feature N-1, where N is a positive integer greater than or equal to 2.

[0168] The derivation relationship is used to describe how a derived feature is generated based on a native feature in the feature derivation process. For example, if post-entry feature 2 is generated based on post-entry feature 1, it is said that post-entry feature 2 has a derivation relationship with post-entry feature 1, and post-entry feature 1 is the native feature, and post-entry feature 2 is the derived feature derived from post-entry feature 1.

[0169] Since the description information can include feature derivation information, the source of the post-entry feature can be determined based on the feature derivation information, that is, the native feature having a derivation relationship with the post-entry feature is determined. Further, the post-entry feature can be managed based on the derivation relationship and the native feature. For example, in actual applications, the post-entry feature can be stored in the form of a data table, so that new features can be generated based on related fields in the data table through feature derivation, and new post-entry features can be obtained through feature entry operations. For example, when it is detected that a post-entry feature has a quality problem, if the post-entry feature is generated based on feature derivation, the original feature of the post-entry feature and the derivation relationship therebetween can be traced back in the feature management system to find the root cause.

[0170] In an embodiment, the terminal can also implement the step of "managing the post-entry feature based on the derivation relationship and the native feature" based on the interaction with the user. Specifically, the step of "managing the post-entry feature based on the derivation relationship and the native feature" can include:

[0171] Based on the derivation relationship, a feature derivation schematic page of the post-entry feature is displayed, where the feature derivation schematic page includes a feature derivation schematic diagram, and the feature derivation schematic diagram includes a feature node corresponding to the native feature and a feature node corresponding to the post-entry feature.

[0172] In response to the node selection operation on the target feature node, the meta information corresponding to the target post-entry feature is displayed on the feature derivation schematic page to perform feature management on the target post-entry feature, where the target post-entry feature is a post-entry feature corresponding to the target node.

[0173] As an example, referring to Figure 6 , the terminal can display a feature derivation schematic page as shown in 10011 based on the derivation relationship between the post-entry feature and the original feature, where the feature derivation schematic page 10011 can include a feature derivation schematic diagram as shown in 10012, and the feature derivation schematic diagram 10012 can include feature nodes corresponding to the original features (i.e., the original features are feature 1, feature 2, and feature 3 in Figure 6 , and the corresponding nodes are node 10013, node 10014, and node 10015, respectively) and feature nodes corresponding to the post-entry features (i.e., the post-entry features are target features in Figure 6 , and the corresponding target nodes are node 10016). The user can select a target feature node in the feature derivation schematic diagram 10012 that he or she wants to understand the meta information, and the terminal can correspondingly display the meta information corresponding to the target post-entry feature on the feature derivation schematic page 10011 in response to the node selection operation on the target feature node to perform feature management on the target post-entry feature, where the target post-entry feature is a post-entry feature corresponding to the target node.

[0174] In actual applications, the description information can further include feature permission information, which can represent the feature access permission of the user group to the post-entry feature. Specifically, the feature management method can further include:

[0175] According to the feature permission information, a target user group corresponding to the post-entry feature is determined, where the target user group includes at least one user, and each user has feature access permission to the post-entry feature;

[0176] The post-entry feature is managed through the target user group.

[0177] The feature permission information is used to represent the feature access permission of the user group to the post-entry feature, and the feature access permission of the post-entry feature refers to the relevant permission of the user to access the post-entry feature.

[0178] The user group refers to a user set composed of at least one user, so if the user cluster has feature access permission to the post-entry feature, each user in the user cluster can have feature access permission to the post-entry feature.

[0179] Therefore, in the present application, a user group having feature access authority for the post-entry feature can be determined according to the feature authority information of the post-entry feature, that is, the target user group corresponding to the post-entry feature, wherein each user in the target user group has feature access authority for the post-entry feature.

[0180] Further, the post-entry feature can be managed by the target user group. For example, the terminal can display a user group management page, wherein the user group management page includes group information of the target user group and an authority management control for the target user group; and in response to a triggering operation on the authority management control, the post-entry feature is managed.

[0181] As an example, the target user group corresponding to the post-entry feature can be a "portrait group", and referring to Figure 7 , the terminal can display a user group management page as shown in 10017, wherein the user group management page 10017 can include group information of the "portrait group", such as group identification number (Identity Document, ID) (i.e. "role ID" in the figure), group name (i.e. "role name" in the figure), group administrator (i.e. "role administrator" in the figure), group creation time, group modification time, and authority management control (specifically including "user management control" and "feature management control" in the figure); the user can trigger the authority management control to implement authority management, and accordingly, the terminal can respond to the triggering operation on the authority management control to manage the post-entry feature, for example, the feature access authority of the target user group for the post-entry feature can be adjusted, thereby managing the post-entry feature; for another example, the users in the target user group can be adjusted to adjust the users having feature access authority, thereby managing the post-entry feature; and the like.

[0182] In an embodiment, the post-entry feature can be managed by adjusting the feature access authority of the target user group for the post-entry feature, specifically, the step of "managing the post-entry feature by the target user group" can include:

[0183] determining a set of accessible features of the target user group, wherein the set of accessible features includes at least one accessible post-entry feature of the target user group, and each user in the target user group has feature access authority for each accessible post-entry feature;

[0184] determining a target post-entry feature to be adjusted from the set of accessible features;

[0185] adjust the feature access right of the target user group based on the target entered feature;

[0186] manage the entered feature based on the adjusted target user group.

[0187] If the target user group has a feature access right for the entered feature, the entered feature is an accessible entered feature of the target user group.

[0188] In this application, there can be multiple ways to determine the accessible feature set of the target user group. For example, the terminal can generate a query request based on the group information of the target user group and send the query request to the server to trigger the server to query the accessible feature set of the target user group based on the query request. For another example, the terminal can record the accessible entered features corresponding to each user group. Therefore, the terminal can query the accessible entered features corresponding to the target user group and determine the accessible feature set of the target user group according to the query result. And so on.

[0189] The adjustment of the entered feature can include multiple cases, such as adjusting the order or priority of the entered feature in the accessible feature set, or deleting the entered feature from the accessible feature set, and so on.

[0190] There can be multiple ways to determine the target entered feature to be adjusted from the accessible feature set, such as receiving a message sent by the server and parsing the message to determine the target entered feature to be adjusted, or determining the target entered feature to be adjusted from the accessible feature set based on the interaction with the user. As an example, the permission management control for the target user group can include a feature management control. The terminal can display a feature adjustment page in response to a trigger operation on the feature management control, where the feature adjustment page includes the accessible feature set of the target user group, the accessible feature set includes at least one feature combination, each feature combination includes at least one accessible entered feature of the target user group, and each feature combination has a corresponding feature adjustment control. The target entered feature to be adjusted is determined from the accessible feature set in response to a trigger operation on the target feature adjustment control.

[0191] For example, the feature management control can be the "feature management" control shown in Figure 7 The user can trigger the feature management control, and the terminal can display the Figure 8The feature adjustment page is shown, wherein the feature adjustment page includes an accessible feature set of the target user group, the accessible feature set includes at least one feature combination, each feature combination includes at least one accessible post-entry feature of the target user group, each feature combination has a corresponding feature adjustment control, such as the space shown in 10018; further, the user can trigger the target feature adjustment control to select the target post-entry feature to be adjusted from the corresponding feature combination, and the terminal can correspondingly determine the target post-entry feature to be adjusted from the accessible feature set in response to the triggering operation of the target feature adjustment control.

[0192] After determining the target post-entry feature to be adjusted, the terminal can further adjust the feature set based on the target post-entry feature, for example, adjusting the order or priority of the target post-entry feature in the accessible feature set; or, for example, deleting the target post-entry feature from the accessible feature set; and the like, so that the feature management of the post-entry feature can be realized by adjusting the feature access permission of the target user group to the post-entry feature, that is, the feature management of the post-entry feature is realized by adjusting the target user group.

[0193] In another embodiment, the adjustment of the user in the target user group can be used to adjust the user with feature access permission, so as to realize the feature management of the post-entry feature. Specifically, the step of "managing the post-entry feature by the target user group" can include:

[0194] Determining the target user to be adjusted from the target user group;

[0195] Adjusting the target user group based on the target user to adjust the target user group;

[0196] The feature management of the post-entry feature is realized by the adjusted target user group.

[0197] Wherein, the adjustment of the user in the user group can include various cases, for example, it can include adjusting the role of the user in the user group; or, for example, it can include deleting the user from the user group; and the like.

[0198] The target user to be adjusted can be determined in various ways from the target user group. For example, the terminal can receive a message sent by the server and parse the message to determine the target user to be adjusted. For another example, the terminal can determine the target user to be adjusted from the target user group based on interaction with the user. As an example, the permission management control for the target user group can include a user management control. In response to a triggering operation on the user management control, the terminal can display a user adjustment page. The user adjustment page includes user information of the users in the target user group and adjustment operation controls for the users. In response to a triggering operation on a target adjustment operation control, the terminal can determine the target user to be adjusted from the target user group.

[0199] For example, the user management control can be a "User Management" control as shown in FIG. 13A. In response to a triggering operation on the user management control, the terminal can display a user adjustment page as shown in FIG. 13B. The user adjustment page includes user information of the users in the target user group and adjustment operation controls for the users (e.g., "Edit" controls and "Operation" controls as shown in FIG. 13B). Figure 7 Figure 9 For example, the user management control can be a "User Management" control as shown in FIG. 13A. In response to a triggering operation on the user management control, the terminal can display a user adjustment page as shown in FIG. 13B. The user adjustment page includes user information of the users in the target user group and adjustment operation controls for the users (e.g., "Edit" controls and "Operation" controls as shown in FIG. 13B). Figure 9

[0200] After determining the target user to be adjusted, the terminal can further adjust the target user group based on the target user to adjust the target user group, for example, to adjust the role of the user in the user group, or to delete the user from the user group, and the like. In this way, the users in the target user group are adjusted to adjust the users with the characteristic access permission, thereby achieving the feature management of the post-entry feature through the post-adjustment target user group.

[0201] As can be seen from the above, the embodiment can obtain feature original data and description information of the feature original data. The description information includes feature attribute information and feature evaluation information. The feature original data is subjected to a feature entry operation according to the description information. The post-entry feature is subjected to feature classification based on the feature attribute information to determine a feature category to which the post-entry feature belongs. The feature evaluation index of the post-entry feature in at least one evaluation dimension is calculated based on the feature evaluation information. The post-entry feature is subjected to feature management according to the feature category and the feature evaluation index.

[0202] ​​The scheme considers multiple links from feature acquisition to application in actual application when managing the feature, for example, the scheme includes the links of obtaining feature raw data, feature entry, feature evaluation, and feature management, which can greatly improve the efficiency of managing the feature. Specifically, the scheme can perform the feature entry operation according to the description information of the feature raw data, which can orderly enter the feature, and facilitate subsequent feature classification, feature query, feature explanation, and the like based on the description information; and the scheme can calculate the feature evaluation index of the entered feature in at least one evaluation dimension, so that the quality of the entered feature can be controlled from multiple aspects, thereby reducing accidents caused by feature quality problems in the feature management process, and improving the robustness of the feature management system.

[0203] In addition, the scheme can be applied to a wide range of business scenarios, for example, the scheme can be applied to feature management of user portrait features, financial modeling features, and features in recommendation scenarios, so that the scheme can be applied to more extensive business scenarios in actual application, and can improve the feature management efficiency from a larger coverage.

[0204] In addition, when the scheme is applied in practice to manage more than 70,000 features, it can support a variety of general feature detection operators (PSI1, PSI6, zero value rate, zero value fluctuation rate, null value rate, null value fluctuation rate, and the like), so that the feature quality monitoring capability reduces the accident rate by more than 90%; compared with the feature engineering scheme in the industry, the scheme is more suitable for financial modeling features, and the feature development time of the feature developer is reduced from 2 hours to 10 minutes, greatly improving the efficiency of the feature developer.

[0205] According to the method described in the above embodiment, the following will be further illustrated in detail by way of example.

[0206] In this embodiment, the feature management device is integrated in the server and the terminal, for example, as shown in Figure 10 The specific process is as follows:

[0207] 201, the terminal obtains feature raw data and description information of the feature raw data, wherein the description information includes feature attribute information and feature evaluation information.

[0208] It is worth noting that in the application, the feature management method can be implemented by cooperation between the front end and the back end. Specifically, the front end can use HyperText Markup Language (HTML), Cascading Style Sheets (CSS), and JavaScript to implement an interactive website, and the back end can use Python, Flask, MySQL, a relational database based on PostgreSQL (i.e. Figure 11 TPG in the application), Elasticsearch, etc. to implement a high-availability supporting stable underlying architecture for data scenarios.

[0209] Among them, MySQL is a relational database management system that uses Structured Query Language (SQL) to access databases, and is the most commonly used database system. Elasticsearch: is a search server based on Lucene, which provides a distributed multi-user full-text search engine, and is an open source enterprise-level search engine developed based on Java language. Elasticsearch is used in cloud computing and can achieve real-time search, stability, reliability, and fast results, and is easy to install and use.

[0210] 202. The terminal performs feature entry operation on the feature raw data according to the description information.

[0211] As an example, the architecture design of the feature management method can be as shown in Figure 11 The data source layer can be used by users to enter features obtained by mining, statistics or crawling, and feature metadata is stored in MySQL. Metadata information is read through MySQL. At the same time, the underlying storage also supports feature synchronization to TPG for portrait analysis. For systems that need to use features, a unified query Application Programming Interface (API) is provided to provide feature retrieval and use.

[0212] 203. The terminal sends the feature attribute information and the feature evaluation information of the entered feature to the server.

[0213] 204. The server classifies the entered feature based on the feature attribute information to determine the feature category to which the entered feature belongs.

[0214] 205. The server sends the classification result of the entered feature to the terminal to enable the terminal to determine the feature category to which the entered feature belongs.

[0215] In the present application, the execution order between steps 204 and 205 is not limited.

[0216] 206、The server calculates the feature evaluation index of the post-entry feature in at least one evaluation dimension based on the feature evaluation information.

[0217] 207、The server sends the feature evaluation index to the terminal.

[0218] 208、The terminal performs feature management on the post-entry feature according to the feature category and the feature evaluation index.

[0219] The terminal can display a corresponding feature management page according to the feature category and the feature evaluation index to perform feature management on the post-entry feature. For example, the terminal can display a feature grouping page as shown in Figure 4 ; for another example, the terminal can display a feature quality monitoring page as shown in Figure 5 .

[0220] In an embodiment, in addition to displaying the feature grouping page according to the feature category, the terminal can also support searching for the post-entry feature. Specifically, the feature grouping page can include a search control, and the terminal can display a search page in response to a triggering operation on the search control, where the search page can include a search condition setting area and a search result display area, the search condition setting area includes at least one search condition setting control; and the terminal displays the search result in the search result display area in response to a search condition setting operation on the search condition setting control.

[0221] As an example, the feature grouping page 1006 can include a search control as shown in 10019, and the user can trigger the search control 10019. Correspondingly, the terminal can display a search page as shown in Figure 12 in response to a triggering operation on the search control 10019, where the search page can include a search condition setting area 20001 and a search result display area 20002, and the search condition setting area includes at least one search condition setting control. The search condition setting control can be used by the user to set search conditions such as keywords, database names, data periods, feature topics, data table names, partition times, feature sources, responsible persons, whether to derive, quality monitoring, etc.

[0222] Further, the user can set the search conditions required for feature search by interacting with the search condition setting control. Correspondingly, the terminal can display the search result in the search result display area in response to a search condition setting operation on the search condition setting control. In this way, the terminal can realize multi-condition combined filtering of features in combination with multiple search conditions.

[0223] After the feature management system is deployed through the foregoing steps, the actual effect can be as shown inFigure 13 As shown in the figure, the metadata module mainly provides data table metadata (including data table partition, data table blood relationship, primary key, partition field, scheduling task, etc. Meta information), which is provided to the feature system when the feature is stored, and the standardized input process is provided. The bottom layer uses a metadata management system to uniformly manage metadata and provide metadata interface capabilities to the outside;

[0224] For the quality monitoring module, when the feature is entered, the feature field configuration monitoring rule (supporting customization) is configured. Due to the large number of data tables and features, the quality monitoring module SQLizes the rules configured by the user, converts the detection logic and threshold into HiveQL, and hands it over to TDW for calculation. If the detection rule detects a quality module exception, the feature developer will be notified of the feature data exception through the alarm center, so as to troubleshoot feature problems (feature blood relationship, alarm details, etc.) through technical means, and timely handle feature problems. At the same time, the quality monitoring module will output feature quality reports through a timing task to ensure feature quality;

[0225] The feature synchronization module synchronizes data in MySQL to TPG, which can enable the use of developed features in portrait analysis and other data analysis scenarios. The portrait analysis system is based on Hermes at the bottom layer, and using TPG can effectively perform analysis;

[0226] Further, in order to serve more upper-layer systems for the use of features, the system provides a unified feature query system, a unified query interface based on Elasticsearch, which provides stable and fast query APIs to query scenarios and external systems through the fast query capability of Elasticsearch, and has good support for financial modeling, deep learning, and strategy scenarios;

[0227] For feature approval and permission management: feature permissions mainly include user permissions for features and user role permissions for features, which aims to solve the problem of users using features. Users can submit an approval form to apply for feature role permissions, and will obtain a batch of feature permissions after approval. It is also part of feature sensitive permission management, mainly suitable for feature permission module management.

[0228] In actual application, for the technical solution process in Figure 13 Lucene or ClickHouse can be used as a substitute, for example, ClickHouse also uses columnar storage structure, supports replica sharding, and supports MergeTree merge tree table at the bottom layer, providing multi-level index operation, which can replace Elasticsearch in massive data aggregation scenarios; For feature quality monitoring, if resources are sufficient, a separate computing task can also be configured, and Hive does not have to be used for data calculation.

[0229] As can be seen from the above, in the feature management, the embodiments of the present application consider various links from production to application of the feature in actual application, for example, the links including obtaining feature original data, feature input, feature evaluation, and feature management, etc., which can greatly improve the efficiency of feature management. Specifically, the scheme can perform feature input operation according to the description information of the feature original data, which can orderly input the feature, and also facilitate subsequent feature classification, feature query, feature explanation, etc. based on the description information; and the scheme can calculate the feature evaluation index of the input feature in at least one evaluation dimension, so that the quality of the input feature can be controlled from multiple aspects, thereby reducing accidents caused by feature quality problems in the feature management process, and improving the robustness of the feature management system.

[0230] In addition, the scheme can be applied to a wide range of business scenarios, for example, the scheme can be applied to feature management of user portrait features, financial modeling features, and features in recommendation scenarios, etc., so that in actual application, the scheme can be applied to more extensive business scenarios, and can improve the feature management efficiency from a larger coverage.

[0231] In addition, the scheme manages the whole life cycle of the feature (demand single→ development→ input system→ effect evaluation→ monitoring→ permission management), solves the problems of difficult management, difficult use, and difficult query of financial features.

[0232] In order to better implement the above method, correspondingly, the embodiments of the present application also provide a feature management device, which can be integrated in a server or a terminal. The server can be a stand-alone physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, and basic cloud computing services such as big data and artificial intelligence platforms. The terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a vehicle-mounted computer, etc., but is not limited thereto.

[0233] For example, as shown in Figure 14 The feature management device can include an obtaining unit 301, an input unit 302, a classification unit 303, a calculation unit 304, and a first management unit 305, as follows:

[0234] The obtaining unit 301 can be used to obtain feature original data and description information of the feature original data, wherein the description information includes feature attribute information and feature evaluation information;

[0235] The input unit 302 can be configured to perform feature inputting on the feature raw data according to the description information;

[0236] The classification unit 303 can be configured to perform feature classification on the inputted feature based on the feature attribute information, to determine a feature category to which the inputted feature belongs;

[0237] The calculation unit 304 can be configured to calculate a feature evaluation index of the inputted feature in at least one evaluation dimension based on the feature evaluation information;

[0238] The first management unit 305 can be configured to perform feature management on the inputted feature according to the feature category and the feature evaluation index.

[0239] In an embodiment, referring to Figure 15 , the feature evaluation information includes a feature evaluation sample set and at least one feature evaluation operator, each of the feature evaluation operators corresponds to one of the evaluation dimensions; and the calculation unit 304 can include:

[0240] The result calculation sub-unit 3041 can be configured to calculate a feature evaluation result of the inputted feature in a corresponding evaluation dimension in the feature evaluation sample set according to the at least one feature evaluation operator;

[0241] The result conversion sub-unit 3042 can be configured to convert the feature evaluation result into the feature evaluation index of the inputted feature in at least one evaluation dimension according to a preset feature evaluation threshold.

[0242] In an embodiment, referring to Figure 16 , the input unit 302 can include:

[0243] The information extraction sub-unit 3021 can be configured to extract meta-information of the feature raw data from the description information;

[0244] The relationship establishment sub-unit 3022 can be configured to establish an association relationship between the meta-information and the feature raw data;

[0245] The feature inputting sub-unit 3023 can be configured to perform feature inputting on the feature raw data based on the association relationship.

[0246] In an embodiment, referring to Figure 17 , the first management unit 305 can include:

[0247] The feature grouping sub-unit 3051 can be configured to perform feature grouping on the inputted feature based on the feature category;

[0248] The combination determining sub-unit 3052 can be configured to determine a feature combination to which the post-entry feature belongs according to the grouping result, wherein the feature combination includes at least one post-entry feature belonging to the same feature category.

[0249] The first management sub-unit 3053 can be configured to perform feature management on the post-entry feature according to the feature combination and the feature evaluation index.

[0250] In an embodiment, referring to Figure 18 The feature attribute information includes feature classification information; and the classification unit 303 can include:

[0251] The candidate determining sub-unit 3031 can be configured to determine a candidate feature category matching the post-entry feature from a preset category set based on the feature classification information.

[0252] The matching calculating sub-unit 3032 can be configured to calculate a matching degree between the post-entry feature and the candidate feature category.

[0253] The category determining sub-unit 3033 can be configured to determine a feature category to which the post-entry feature belongs according to the matching degree.

[0254] In an embodiment, referring to Figure 19 The description information further includes feature derivation information, the feature derivation information indicating that the post-entry feature is generated based on feature derivation, and the feature management apparatus further includes:

[0255] The original determining unit 306 can be configured to determine an original feature having a derivation relationship with the post-entry feature according to the feature derivation information, wherein the post-entry feature is generated based on the original feature.

[0256] The second management unit 307 can be configured to perform feature management on the post-entry feature based on the derivation relationship and the original feature.

[0257] In an embodiment, referring to Figure 20 The second management unit 307 can include:

[0258] The page displaying sub-unit 3071 can be configured to display a feature derivation schematic page of the post-entry feature based on the derivation relationship, wherein the feature derivation schematic page includes a feature derivation schematic diagram including a feature node corresponding to the original feature and a feature node corresponding to the post-entry feature.

[0259] The information display sub-unit 3072 can be configured to display meta information corresponding to the target post-entry feature in the feature derivation schematic page in response to a node selection operation on the target feature node, so as to perform feature management on the target post-entry feature, where the target post-entry feature is a post-entry feature corresponding to the target node.

[0260] In an embodiment, referring to Figure 21 , the description information further comprises feature permission information, the feature permission information representing feature access permissions of a user group to the post-entry feature, and the feature management apparatus further comprises:

[0261] The group determination unit 308 can be configured to determine a target user group corresponding to the post-entry feature according to the feature permission information, where the target user group comprises at least one user, and each of the users has feature access permissions to the post-entry feature.

[0262] The third management unit 309 can be configured to perform feature management on the post-entry feature by using the target user group.

[0263] In an embodiment, referring to Figure 22 , the third management unit 309 can comprise:

[0264] The set determination sub-unit 3091 can be configured to determine an accessible feature set of the target user group, where the accessible feature set comprises at least one accessible post-entry feature of the target user group, and each of the users in the target user group has feature access permissions to each of the accessible post-entry features.

[0265] The feature determination sub-unit 3092 can be configured to determine a target post-entry feature to be adjusted from the accessible feature set.

[0266] The feature adjustment sub-unit 3093 can be configured to perform feature adjustment on the accessible feature set based on the target post-entry feature, so as to adjust feature access permissions of the target user group.

[0267] The second management sub-unit 3094 can be configured to perform feature management on the post-entry feature by using an adjusted target user group.

[0268] In an embodiment, referring to Figure 23 , the third management unit 309 can comprise:

[0269] The user determination sub-unit 3095 can be configured to determine a target user to be adjusted from the target user group.

[0270] The user adjustment sub-unit 3096 can be configured to perform user adjustment on the target user group based on the target user, so as to adjust the target user group.

[0271] The third management sub-unit 3097 can be configured to perform feature management on the recorded feature based on the adjusted target user group.

[0272] In specific implementation, the above units can be implemented as independent entities, or can be combined as the same or several entities. The specific implementation of the above units can refer to the method embodiments above, and will not be described here.

[0273] As can be seen, in the feature management apparatus, the acquisition unit 301 is configured to acquire feature raw data and description information of the feature raw data, wherein the description information comprises feature attribute information and feature evaluation information; the recording unit 302 is configured to perform feature recording operation on the feature raw data based on the description information; the classification unit 303 is configured to perform feature classification on the recorded feature based on the feature attribute information, so as to determine the feature category to which the recorded feature belongs; the calculation unit 304 is configured to calculate feature evaluation indexes of the recorded feature in at least one evaluation dimension based on the feature evaluation information; and the first management unit 305 is configured to perform feature management on the recorded feature based on the feature category and the feature evaluation indexes.

[0274] The scheme considers multiple links from production to application of features in actual application when managing the features, for example, including acquisition of feature raw data, feature recording, feature evaluation, and feature management, which can greatly improve the efficiency of feature management. Specifically, the scheme can perform feature recording operation based on the description information of the feature raw data, which can orderly record the features, and facilitate subsequent feature classification, feature query, feature explanation, etc. based on the description information; and the scheme can calculate feature evaluation indexes of the recorded feature in at least one evaluation dimension, so as to control the quality of the recorded feature from multiple aspects, thereby reducing accidents caused by feature quality problems in the feature management process, and improving the robustness of the feature management system.

[0275] In addition, the scheme can be applied to a wide range of business scenarios, for example, the scheme can be applied to feature management of user portrait features, financial modeling features, and features in recommendation scenarios, so that the scheme can be applied to more extensive business scenarios in actual application, and can improve the feature management efficiency from a larger coverage.

[0276] In addition, the embodiment of the present application also provides a computer device, which can be a server or terminal device, etc. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The terminal can be a smart phone, tablet computer, laptop computer, desktop computer, smart speaker, smart watch, car computer, etc., but is not limited to this. Figure 24 , which shows a schematic diagram of the structure of the computer device involved in the embodiment of the present application, specifically:

[0277] The computer device may include a memory 401 having one or more computer-readable storage media, an input unit 402, a processor 403 having one or more processing cores, and a power supply 404. It will be understood by those skilled in the art that Figure 24 The computer device structure shown in the figure does not constitute a limitation on the computer device, and may include Figure 24 More or fewer components, or combinations of certain components, or different arrangements of components may be shown.

[0278] The memory 401 can be used to store software programs and modules, and the processor 403 executes various functional applications and data processing by running the software programs and modules stored in the memory 401. The memory 401 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the computer device (such as audio data, a phone book, etc.). In addition, the memory 401 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage device. Accordingly, the memory 401 may also include a memory controller to provide the processor 403 and the input unit 402 with access to the memory 401.

[0279] The input unit 402 can be configured to receive input of digital or character information, and generate a keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control. Specifically, in one specific embodiment, the input unit 402 can include a touch-sensitive surface and other input devices. The touch-sensitive surface, also known as a touch display or touchpad, can collect touch operations of a user thereon or adjacent thereto (such as operations of a user using a finger, a stylus, or any suitable object or accessory on or adjacent to the touch-sensitive surface), and drive corresponding connection devices according to a pre-set program. Optionally, the touch-sensitive surface can include two parts of a touch detection device and a touch controller. Among them, the touch detection device detects the touch position of the user and detects the signal brought by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device, and converts it into touch coordinates and sends it to the processor 403, and can also receive commands from the processor 403 and execute them. In addition, the touch-sensitive surface can be implemented in various types such as resistive, capacitive, infrared, and surface acoustic wave. In addition to the touch-sensitive surface, the input unit 402 can also include other input devices. Specifically, the other input devices can include one or more of a physical keyboard, function keys (such as volume control keys, on-off keys, etc.), trackballs, mice, joysticks, etc.

[0280] The processor 403 is the control center of the computer device, which connects all parts of the mobile phone through various interfaces and lines, executes various functions of the computer device and processes data by running or executing software programs and / or modules stored in the memory 401 and calling data stored in the memory 401, thereby monitoring the mobile phone as a whole. Optionally, the processor 403 can include one or more processing cores; preferably, the processor 403 can integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface and application programs, and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor can also not be integrated into the processor 403.

[0281] The computer device further includes a power supply 404 (such as a battery) for supplying power to various components, and preferably the power supply can be logically connected to the processor 403 through a power management system, so as to realize functions such as management of charging, discharging and power consumption management through the power management system. The power supply 404 can also include one or more than one direct or alternating current power supply, a recharging system, a power failure detection circuit, a power converter or inverter, a power state indicator, and any other components.

[0282] Although not shown, the computer device can further include a camera, a Bluetooth module, etc., which will not be described herein. Specifically in the present embodiment, the processor 403 in the computer device will load one or more executable files corresponding to processes of one or more application programs into the memory 401 according to the following instructions, and run the application programs stored in the memory 401 by the processor 403, so as to realize various functions, as follows:

[0283] obtaining feature raw data and description information of the feature raw data, wherein the description information comprises feature attribute information and feature evaluation information; performing a feature entry operation on the feature raw data according to the description information; performing feature classification on the entered feature based on the feature attribute information, to determine a feature category to which the entered feature belongs; calculating a feature evaluation index of the entered feature in at least one evaluation dimension based on the feature evaluation information; and performing feature management on the entered feature according to the feature category and the feature evaluation index.

[0284] The specific implementation of each operation above can refer to the foregoing embodiments, which will not be described herein.

[0285] As can be seen from the above, the computer device of the present embodiment considers multiple links from production to application of features in actual application when managing the features, for example, which includes links such as obtaining feature raw data, feature entry, feature evaluation, and feature management, which can greatly improve the efficiency of managing the features. Specifically, the scheme can perform a feature entry operation according to the description information of the feature raw data, which can orderly enter the features, and also facilitates subsequent feature classification, feature query, feature explanation, etc. based on the description information; and the scheme can calculate a feature evaluation index of the entered feature in at least one evaluation dimension, so that the quality of the entered feature can be controlled from multiple aspects, thereby reducing accidents caused by feature quality problems in the feature management process, and improving the robustness of the feature management system.

[0286] In addition, the scheme can be applicable to a wide range of business scenarios, for example, the scheme can be applicable to feature management of user portrait features, financial modeling features, and features in a recommendation scenario, so that in actual application, the scheme can be applicable to more extensive business scenarios, and can improve the feature management efficiency from a larger coverage.

[0287] Those skilled in the art can understand that all or part of the steps of the various methods of the above embodiments can be completed by instructions, or by related hardware controlled by the instructions, which can be stored in a computer readable storage medium and loaded and executed by a processor.

[0288] To this end, an embodiment of the present application provides a storage medium having stored therein a plurality of instructions capable of being loaded by a processor to execute steps in any of the feature management methods provided by the embodiments of the present application. For example, the instructions can execute the following steps:

[0289] obtaining feature raw data and description information of the feature raw data, wherein the description information comprises feature attribute information and feature evaluation information; performing a feature entry operation on the feature raw data according to the description information; performing feature classification on the entered feature based on the feature attribute information to determine a feature category to which the entered feature belongs; calculating a feature evaluation index of the entered feature in at least one evaluation dimension based on the feature evaluation information; and performing feature management on the entered feature according to the feature category and the feature evaluation index.

[0290] The specific implementation of the above operations can be referred to the foregoing embodiments, which will not be described here again.

[0291] The storage medium can include a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0292] Since the instructions stored in the storage medium can execute steps in any of the feature management methods provided by the embodiments of the present application, the beneficial effects that can be achieved by any of the feature management methods provided by the embodiments of the present application can be achieved, which will not be described here again.

[0293] According to an aspect of the present application, a computer program product or computer program is provided, which includes computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to enable the computer device to perform the method provided in any of the various optional implementations of the feature management aspect.

[0294] According to an aspect of the present application, a computer program product or computer program is provided, which includes computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to enable the computer device to perform the method provided in any of the various optional implementations of the feature management aspect.

[0295] The above describes in detail a feature management method, device, computer equipment, storage medium and product provided by the embodiments of the present application. The principles and implementation manners of the present application are described by applying specific examples. The above description of the embodiments is only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manners and application ranges will be changed. In conclusion, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A feature management method, comprising: obtaining feature raw data and description information of the feature raw data, wherein the description information comprises feature attribute information, feature evaluation information and feature derivation information; performing feature entry operation on the feature raw data according to the description information; performing feature classification on the entered feature based on the feature attribute information to determine a feature category to which the entered feature belongs; calculating a feature evaluation index of the entered feature in at least one evaluation dimension based on the feature evaluation information; performing feature management on the entered feature according to the feature category and the feature evaluation index; determining a native feature having a derivation relationship with the entered feature according to the feature derivation information, wherein the entered feature is derived from the native feature; performing feature management on the entered feature based on the derivation relationship and the native feature, comprising: displaying a feature derivation schematic page of the entered feature based on the derivation relationship, wherein the feature derivation schematic page comprises a feature derivation schematic diagram comprising a feature node corresponding to the native feature, a feature node corresponding to the entered feature, and a derivation relationship between the native feature and the entered feature; in response to a node selection operation on a target node, displaying meta information corresponding to a target entered feature on the feature derivation schematic page to perform feature management on the target entered feature, wherein the target node is the feature node corresponding to the entered feature, and the target entered feature is the entered feature.

2. The feature management method according to claim 1, characterized in that, The feature evaluation information comprises a feature evaluation sample set and at least one feature evaluation operator, and each feature evaluation operator corresponds to one evaluation dimension. The calculation of the feature evaluation index of the entered feature in at least one evaluation dimension based on the feature evaluation information comprises: calculating a feature evaluation result of the entered feature in a corresponding evaluation dimension in the feature evaluation sample set according to the at least one feature evaluation operator; converting the feature evaluation result into the feature evaluation index of the entered feature in at least one evaluation dimension according to a preset feature evaluation threshold.

3. The feature management method of claim 1, wherein, The feature entry operation on the feature raw data according to the description information comprises: extracting meta information of the feature raw data from the description information; establishing an association relationship between the meta information and the feature raw data; performing feature entry operation on the feature raw data based on the association relationship.

4. The feature management method of claim 1, wherein, The feature management on the entered feature according to the feature category and the feature evaluation index comprises: performing feature grouping on the entered feature based on the feature category; determining a feature combination to which the entered feature belongs according to the grouping result, wherein the feature combination comprises at least one entered feature belonging to the same feature category; performing feature management on the entered feature according to the feature combination and the feature evaluation index.

5. The feature management method of claim 1, wherein, The feature attribute information comprises feature classification information. Based on the feature attribute information, the post-entry feature is classified to determine a feature category to which the post-entry feature belongs, including: Based on the feature classification information, a candidate feature category matching the post-entry feature is determined from a preset category set; A matching degree between the post-entry feature and the candidate feature category is calculated; According to the matching degree, a feature category to which the post-entry feature belongs is determined.

6. The feature management method of claim 1, wherein, The description information further includes feature permission information, and the feature permission information represents feature access permissions of a user group to the post-entry feature. The method further includes: According to the feature permission information, a target user group corresponding to the post-entry feature is determined, wherein the target user group includes at least one user, and each user has feature access permissions to the post-entry feature; The post-entry feature is managed through the target user group.

7. The feature management method of claim 6, wherein, The post-entry feature is managed through the target user group, including: An accessible feature set of the target user group is determined, wherein the accessible feature set includes at least one accessible post-entry feature of the target user group, and a user in the target user group has feature access permissions to each accessible post-entry feature; A target post-entry feature to be adjusted is determined from the accessible feature set; Based on the target post-entry feature, the accessible feature set is adjusted to adjust the feature access permissions of the target user group; The post-entry feature is managed through an adjusted target user group.

8. The feature management method of claim 6, wherein, The post-entry feature is managed through the target user group, including: A target user to be adjusted is determined from the target user group; Based on the target user, the target user group is adjusted to adjust the target user group; The post-entry feature is managed through an adjusted target user group.

9. A feature management apparatus, characterized by, Including: An acquisition unit is configured to acquire feature original data and description information of the feature original data, wherein the description information includes feature attribute information, feature evaluation information, and feature derivation information; An entry unit is configured to perform a feature entry operation on the feature original data according to the description information; A classification unit is configured to classify a post-entry feature based on the feature attribute information to determine a feature category to which the post-entry feature belongs; A calculation unit is configured to calculate a feature evaluation index of the post-entry feature in at least one evaluation dimension based on the feature evaluation information; A first management unit is configured to manage the post-entry feature according to the feature category and the feature evaluation index; The apparatus is further configured to determine a native feature having a derivation relationship with the post-entry feature based on the feature derivation information, wherein the post-entry feature is derived and generated based on the native feature; The post-entry feature is managed based on the derivation relationship and the native feature, including: based on the derived relationship, a feature derivation schematic page of the post-entry feature is displayed, wherein the feature derivation schematic page includes a feature derivation schematic diagram including a feature node corresponding to the native feature, a feature node corresponding to the post-entry feature, and a derived relationship between the native feature and the post-entry feature; in response to a node selection operation on a target node, meta information corresponding to a target post-entry feature is displayed on the feature derivation schematic page to perform feature management on the target post-entry feature, wherein the target node is a feature node corresponding to the post-entry feature, and the target post-entry feature is the post-entry feature.

10. An electronic device, comprising: comprising a memory and a processor; the memory stores an application program, and the processor is configured to run the application program in the memory to perform the operations in the feature management method of any one of claims 1 to 8.

11. A storage medium, characterized by The storage medium stores a plurality of instructions, which are adapted to be loaded by a processor to perform the steps in the feature management of any one of claims 1 to 8.

12. A computer program product comprising computer programs / instructions, characterized in that, The computer program / instructions, when executed by the processor, implement the steps in the feature management method of any one of claims 1 to 8. The computer program / instructions, when executed by the processor, implement the steps in the feature management method of any one of claims 1 to 8.

Images