CN117807282A

CN117807282A - Service data processing method, device, electronic equipment and readable storage medium

Info

Publication number: CN117807282A
Application number: CN202410232663.XA
Authority: CN
Inventors: 刘晨; 吴昊; 郑保卫; 李勇; 敖劲松
Original assignee: Encore Beijing Information Technology Co ltd
Current assignee: Encore Beijing Information Technology Co ltd
Priority date: 2024-03-01
Filing date: 2024-03-01
Publication date: 2024-04-02

Abstract

The application provides a business data processing method, a business data processing device, electronic equipment and a readable storage medium, and relates to the field of data processing. In the method, firstly, retrieval asset data is acquired; responding to the selection operation of a preset target marking model, marking the retrieved asset data by using the target marking model, and obtaining a marking result; according to the marking result, carrying out first recommendation by using a preset recommendation strategy to obtain a related recommendation result, wherein the related recommendation result comprises recommendation data and labels corresponding to the recommendation data; acquiring a preset asset marking list, wherein the asset marking list comprises a plurality of asset data to be marked and marking labels corresponding to the asset data to be marked; obtaining a matching result according to the recommended data, the label corresponding to the recommended data, the asset data to be marked and the marking label; and performing second recommendation according to the matching result and the retrieved asset data to obtain recommended asset data, and recommending accurate asset data for the user.

Description

Service data processing method, device, electronic equipment and readable storage medium

Technical Field

The present invention relates to the field of data processing technologies, and in particular, to a service data processing method, a device, an electronic device, and a readable storage medium.

Background

Asset data of an enterprise is data that the enterprise legally owns or controls, and is generally classified into different categories according to different properties, for example, different asset data such as products, markets, orders, costs, incomes, services, channels, and the like can be classified according to attributes. The asset data checking and intelligent marking can directly or indirectly bring economic benefit, new development opportunities are provided for enterprises, intelligent recommendation is an application scene of the asset data checking and intelligent marking of the enterprises, and suitable asset data are recommended for users, so that the enterprise asset value is realized. Intelligent recommendation recommends recommended content wanted by a user by inputting query data by the user, however, when the user queries the data, the result of the search is not very clear, and usually some related content is input, but the related content is fuzzy, so that accurate asset data cannot be recommended for the user.

Disclosure of Invention

The application provides a business data processing method, a business data processing device, electronic equipment and a readable storage medium, which can recommend accurate asset data for users.

The technical scheme of the embodiment of the application is as follows:

in a first aspect, an embodiment of the present application provides a service data processing method, where the method includes:

Acquiring retrieval asset data;

responding to the selection operation of a preset target marking model, and marking the retrieved asset data by using the target marking model to obtain a marking result;

according to the marking result, performing first recommendation by using a preset recommendation strategy to obtain a related recommendation result, wherein the related recommendation result comprises recommendation data and labels corresponding to the recommendation data;

acquiring a preset asset marking list, wherein the asset marking list comprises a plurality of asset data to be marked and marking labels corresponding to the asset data to be marked;

obtaining a matching result according to the recommended data, the label corresponding to the recommended data, the asset data to be marked and the marking label;

and performing second recommendation according to the matching result and the retrieval asset data to obtain recommended asset data.

In the technical scheme, firstly, the search asset data is obtained, the search asset data is search content input by a user, and the search asset data is obtained to provide data support for recommending accurate asset data to the user later; in response to the selection operation of a preset target marking model, marking the retrieved asset data by using the target marking model to obtain a marking result, and marking the retrieved asset data to obtain a marking result indicating the category of the retrieved asset data, thereby being beneficial to accurate recommendation; according to the marking result, a preset recommendation strategy is utilized to conduct first recommendation, a related recommendation result is obtained, the related recommendation result comprises recommendation data and labels corresponding to the recommendation data, and firstly, the recommendation strategy is utilized to conduct rough recommendation, so that recommendation accuracy is improved; acquiring a preset asset marking list, wherein the asset marking list comprises a plurality of asset data to be marked and marking labels corresponding to the asset data to be marked, and providing data support for follow-up accurate recommendation; obtaining a matching result according to the recommended data, the label corresponding to the recommended data, the asset data to be marked and the marking label; and performing second recommendation according to the matching result and the retrieved asset data to obtain recommended asset data, and performing the second recommendation can obtain more accurate recommended asset data.

In some embodiments of the present application, before the obtaining the preset asset marking manifest, the method further includes:

acquiring a plurality of asset data to be marked, and preprocessing each asset data to be marked to obtain a plurality of preprocessed asset data;

marking each piece of preprocessed asset data by using the target marking model to obtain marking labels;

and obtaining the asset marking list according to the asset data to be marked and marking labels corresponding to the asset data to be marked.

According to the technical scheme, the asset data to be marked is preprocessed, namely data screening and data cleaning are performed, so that the calculated amount can be reduced, and the marking accuracy can be improved. And then, marking the preprocessed asset data by using a target marking model to obtain marking labels, and forming an asset marking list by the asset data to be marked and the marking labels, so that the asset marking list is favorable for matching according to the asset marking list in the follow-up process so as to recommend accurate asset data.

In some embodiments of the present application, before said marking each of said preprocessed asset data with said target marking model, the method further comprises:

Acquiring an asset sample data set, wherein the asset sample data set comprises a plurality of asset sample data and marking labels corresponding to the asset sample data;

preprocessing each asset sample data to obtain a plurality of preprocessed sample data;

performing first sample equalization processing on each piece of preprocessed sample data to obtain a training data set;

performing second sample equalization processing on each piece of preprocessed sample data to obtain a verification data set;

adjusting parameters of a preset first initial marking model by using the training data set to obtain a second initial marking model;

and verifying the second initial marking model by using the verification data set, and obtaining the target marking model under the condition that the accuracy is larger than a preset accuracy threshold.

According to the technical scheme, the generalization of the model can be improved by preprocessing the asset sample data set, namely, performing data screening and data cleaning. And performing sample equalization processing to obtain a training set and a verification set, performing parameter adjustment by using the training set, and performing verification by using the verification set to obtain a target marking model, wherein the target marking model has better generalization and robustness.

In some embodiments of the present application, the performing a first sample equalization process on each of the preprocessed sample data to obtain a training data set includes:

performing duplication removal processing on the preprocessed sample data according to the category, the field Chinese name and the field English name to obtain first duplication-removed sample data;

selecting N categories with the number being arranged in front from the first de-duplicated sample data to obtain selected sample data, wherein N is an integer greater than or equal to 2;

performing deduplication on the selected sample data according to the category and the field Chinese name to obtain second deduplication sample data;

removing categories with the number not exceeding the preset number in the second duplicate removal sample data to obtain the training data set;

and performing second sample equalization processing on each piece of preprocessed sample data to obtain a verification data set, wherein the second sample equalization processing comprises the following steps:

and removing categories of which the number is not more than a preset number in the preprocessed sample data to obtain the verification data set.

In the technical scheme, the duplication elimination is performed according to the category, the field Chinese name and the field English name, and then the duplication elimination is performed according to the category and the field Chinese name, so that the effective duplication elimination is realized, and the repeated record is avoided. And the influence of invalid data on the model accuracy is reduced by removing the categories of which the number is not more than the preset number in the data.

In some embodiments of the present application, before the marking the retrieved asset data with the target marking model, the method further includes:

vector processing is carried out on the retrieved asset data to obtain an asset data vector;

performing data dimension-increasing processing on the retrieved asset data to obtain a multidimensional data vector;

and carrying out fusion processing on the asset data vector and the multi-dimensional data vector, and then inputting the asset data vector and the multi-dimensional data vector into the target marking model.

According to the technical scheme, the associated data of the retrieved asset data can be increased by carrying out data upgrading maintenance so as to clearly recommend the target, and then data fusion is carried out, so that the representation capability of the data is increased, and the accuracy of model output is improved.

In some embodiments of the present application, the obtaining a matching result according to the recommended data, the tag corresponding to the recommended data, the asset data to be marked, and the marking tag includes:

and classifying the label corresponding to the recommended data and the marking label, performing similarity calculation on the recommended data and the asset data to be marked, and screening out the data with consistent label classification and similarity smaller than a preset threshold value to obtain the matching result.

In the technical scheme, accurate recommendation is performed according to the category and the similarity, so that a matching result is obtained, and accurate asset data is recommended for a user according to the matching result.

In some embodiments of the present application, the recommended policy is at least one of a quantity limit, a matching degree limit, and a combination limit.

In the technical scheme, the quantity limitation, the matching degree limitation and the combination limitation have different advantages, and a user can select corresponding strategy recommendation according to requirements, so that the recommended content meets the requirements of the user.

In a second aspect, an embodiment of the present application provides a service data processing apparatus, where the apparatus includes:

the first data acquisition module is used for acquiring the retrieval asset data;

the marking module is used for responding to the selection operation of a preset target marking model, and marking the retrieved asset data by using the target marking model to obtain a marking result;

the first recommendation module is used for carrying out first recommendation by utilizing a preset recommendation strategy according to the marking result to obtain a related recommendation result, wherein the related recommendation result comprises recommendation data and labels corresponding to the recommendation data;

the second data acquisition module is used for acquiring a preset asset marking list, wherein the asset marking list comprises a plurality of asset data to be marked and marking labels corresponding to the asset data to be marked;

The matching module is used for obtaining a matching result according to the recommended data, the label corresponding to the recommended data, the asset data to be marked and the marking label;

and the second recommending module is used for carrying out second recommendation according to the matching result and the retrieval asset data to obtain recommended asset data.

In a third aspect, an embodiment of the present application provides an electronic device, including a processor, a memory, a user interface, and a network interface, where the memory is configured to store instructions, and the user interface and the network interface are configured to communicate with other devices, and the processor is configured to execute the instructions stored in the memory, so that the electronic device performs the method provided in any one of the first aspect above.

In a fourth aspect, embodiments of the present application provide a computer-readable storage medium storing instructions that, when executed, perform the method of any one of the first aspects provided above.

In summary, one or more technical solutions provided in the embodiments of the present application at least have the following technical effects or advantages:

1. the technical means that the retrieved asset data is obtained, the retrieved asset data is marked by utilizing the target marking model, the first recommendation is carried out according to the marking result to obtain the related recommendation result, so that the asset data similar to the retrieved asset data is screened out, the matching result is obtained according to the related recommendation result and the asset marking list, the second recommendation is carried out according to the matching result and the retrieved asset data, and the recommended asset data is obtained is adopted, and therefore the problem that the accurate asset data cannot be recommended for a user due to the fact that the input related content is fuzzy in the related technology is effectively solved. And marking the retrieved asset data through the target marking model, carrying out first recommendation according to the marking result, screening out asset data similar to the retrieved asset data, carrying out second recommendation according to the asset marking list, and recommending accurate asset data for a user through multi-layer recommendation.

2. The data is processed by adopting data preprocessing and sample equalization, so that generalization and robustness of the model can be improved.

3. Vector processing and data dimension increasing processing are carried out on the data, and the data are input into the model, so that the accuracy of the training model can be improved.

Drawings

FIG. 1 is a flow chart of a method for processing business data according to an embodiment of the present application;

FIG. 2 is a second flow chart of a business data processing method according to an embodiment of the present application;

FIG. 3 is a third flow chart of a business data processing method according to an embodiment of the present application;

FIG. 4 is a flow chart illustrating a method for processing business data according to an embodiment of the present application;

fig. 5 is a schematic diagram of label division of a service data processing method according to an embodiment of the present application;

FIG. 6 is a flow chart illustrating a sub-step of step S930 of FIG. 4;

FIG. 7 is a sample equalization schematic of a method for processing service data according to an embodiment of the present application;

FIG. 8 is a schematic overall flow chart of a business data processing method according to an embodiment of the present application;

FIG. 9 is a schematic structural diagram of a service data processing device according to an embodiment of the present application;

Fig. 10 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

In order to make those skilled in the art better understand the technical solutions in the present specification, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is obvious that the described embodiments are only some embodiments of the present application, but not all embodiments.

In the description of embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described herein as "such as" or "for example" should not be construed as preferred or advantageous over other embodiments or designs. Rather, the use of words such as "or" for example "is intended to present related concepts in a concrete fashion.

In the description of the embodiments of the present application, the term "plurality" means two or more. For example, a plurality of systems means two or more systems, and a plurality of screen terminals means two or more screen terminals. Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating an indicated technical feature. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless expressly specified otherwise.

The embodiment of the application provides a business data processing method, a device, electronic equipment and a readable storage medium, wherein the business data processing method firstly acquires retrieval asset data, the retrieval asset data is retrieval content input by a user, and the retrieval asset data is acquired to provide data support for recommending accurate asset data to the user later; in response to the selection operation of a preset target marking model, marking the retrieved asset data by using the target marking model to obtain a marking result, and marking the retrieved asset data to obtain a marking result indicating the category of the retrieved asset data, thereby being beneficial to accurate recommendation; according to the marking result, a preset recommendation strategy is utilized to conduct first recommendation, a related recommendation result is obtained, the related recommendation result comprises recommendation data and labels corresponding to the recommendation data, and firstly, the recommendation strategy is utilized to conduct rough recommendation, so that recommendation accuracy is improved; acquiring a preset asset marking list, wherein the asset marking list comprises a plurality of asset data to be marked and marking labels corresponding to the asset data to be marked, and providing data support for follow-up accurate recommendation; obtaining a matching result according to the recommended data, the label corresponding to the recommended data, the asset data to be marked and the marking label; and performing second recommendation according to the matching result and the retrieved asset data to obtain recommended asset data, and performing the second recommendation can obtain more accurate recommended asset data. Compared with the prior art that the input associated content is fuzzy and accurate asset data cannot be recommended to the user, the method and the device for recommending the asset data in the multi-layer mode are capable of marking the retrieved asset data through the target marking model, performing first recommendation according to the marking result, screening out the asset data similar to the retrieved asset data, performing second recommendation according to the asset marking list, and recommending the accurate asset data to the user through the multi-layer recommendation.

It should be noted that, the business data processing method is applied to the fields of banking, commercial asset recommendation and the like, and can also be applied to other asset recommendation fields. Firstly, through asset inventory marking management, the sorted asset marking list is used for intelligent recommendation, so that accurate asset data can be recommended for users.

The technical scheme provided by the embodiment of the application is further described below by taking banking industry as an example with reference to the accompanying drawings.

Referring to fig. 1, fig. 1 is a flow chart of a service data processing method according to an embodiment of the present application. The service data processing method is applied to a service data processing apparatus, and is executed by a processor in an electronic device or a readable storage medium, and includes steps S100, S200, S300, S400, S500, and S600.

Step S100, obtaining retrieval asset data.

In an embodiment, the search property data is a brief summary or a specific description of data that the user wants to obtain, and the search property data may include loan policy search, financial search of banking, deposit policy search, etc., and the user may input the search property data in a preset webpage or APP, and obtain the search property data through a preset interface, so as to facilitate the subsequent recommendation according to the search property data and provide data support. Wherein the preset interface includes an input () function.

Step S200, in response to the selection operation of the preset target marking model, marking the retrieved asset data by using the target marking model to obtain a marking result.

In an embodiment, the user selects the target marking model on the webpage, and selects different model recommendation modes, so that the recommendation data can be more accurate by the selected target marking model. And responding to the selection operation of a preset target marking model, and marking the retrieved asset data by using the target marking model to obtain a marking result. The marking result is that the data is divided according to the attribute or the category, and the obtained specific subclass is divided according to the industry division standard. Because the target marking model is trained in advance, the category of the retrieved asset data can be accurately marked by adopting the target marking model, and the follow-up recommendation according to the marking result is facilitated. The target marking model can be a bert model, a variant of the bert model, and other deep learning models, and can meet the accurate marking requirement, and details are omitted here.

As shown in fig. 2, before the target marking model is used to mark the retrieved asset data, the business data processing method further includes, but is not limited to, the following steps:

Step S710, vector processing is performed on the retrieved asset data to obtain an asset data vector.

In one embodiment, before the retrieved asset data is input into the target marking model, vector processing is performed on the retrieved asset data, specifically: and vectorizing the retrieved asset data by using a preset data characterization algorithm to obtain an asset data vector. Through data vectorization, the search asset data is converted into the data types which can be processed by the target marking model, and the calculation efficiency is improved.

Step S720, performing data dimension-lifting processing on the retrieved asset data to obtain a multidimensional data vector.

In an embodiment, the data dimension increase is related data expansion for the data, as shown in fig. 8, the data dimension increase for the user data can be expanded into roles, departments, attention businesses, history retrieval records, jurisdictional systems and the like, and the data dimension increase process can be performed for the retrieval asset data according to a pre-established knowledge graph to obtain a multidimensional data vector. The search range of the retrieved asset data can be improved by carrying out data dimension-increasing processing, and the recommendation accuracy is improved. The knowledge graph can reflect the association relation of each data, so that the knowledge graph is utilized to perform rapid data dimension increase, the knowledge graph is constructed in advance according to a specific service scene, and the construction process is similar to that of other knowledge graphs, and is not repeated here.

And step S730, the asset data vector and the multi-dimensional data vector are fused and then input into a target marking model.

In an embodiment, the asset data vector and the multidimensional data vector are fused, the asset data vector and the multidimensional data vector can be adopted for characteristic connection, then the asset data vector and the multidimensional data vector are input into a target marking model, the target marking model is used for marking, and a more accurate marking category can be output, so that the recommendation accuracy is improved.

Step S300, according to the marking result, performing first recommendation by using a preset recommendation strategy to obtain a related recommendation result, wherein the related recommendation result comprises recommendation data and labels corresponding to the recommendation data.

In one embodiment, the first recommendation is a rough recommendation based on retrieving the asset data, the data of a different category than the tag is removed first, and only the content related to retrieving the asset data is recommended. And (3) according to the marking result obtained in the step (S200), carrying out first recommendation by using a preset recommendation strategy to obtain a related recommendation result, and facilitating the follow-up promotion according to the related recommendation result, so as to recommend accurate asset data for a user. The related recommendation results comprise recommendation data and labels corresponding to the recommendation data, the labels corresponding to the recommendation data and the marking results belong to the same category, the recommendation data and the retrieval asset data have strong correlation, the correlation is represented by the matching degree, and the higher the matching degree is, the stronger the correlation is represented.

In another embodiment, the recommended policy is at least one of a quantity limit, a matching degree limit, a combination limit. Under the condition that the recommendation strategy is limited in number, recommending the data of the first a of the matching degree arrangement, and recommending the recommendation strategy with higher recommendation speed. Wherein a is an integer greater than 0 and less than 100, a may be 5, etc., and may be adjusted according to circumstances, which will not be described herein. Under the condition that the recommendation strategy is the matching degree limitation, recommending data with the matching degree larger than a preset association threshold value, wherein the recommendation strategy is faster in recommendation speed and higher in recommendation accuracy. The association threshold may be 70%, etc., and may be adjusted according to circumstances, which will not be described herein. Under the condition that the recommendation strategy is limited by combination, the recommendation matching degree is larger than a preset association threshold value, and the matching degree arranges the first a pieces of data, so that the recommendation strategy has slower recommendation speed and highest recommendation accuracy. Illustratively, the a pieces of data of the default matching degree arrangement are 10 pieces, the association threshold is 70%, and then different strategies are as follows: the number is limited to be 5 before the matching degree is arranged, the matching degree is limited to be more than 70%, and the combination limit is more than or equal to 70% and the matching degree is the highest 5 before.

Step S400, a preset asset marking list is obtained, wherein the asset marking list comprises a plurality of asset data to be marked and marking labels corresponding to the asset data to be marked.

In an embodiment, the preset asset marking list is a record of the asset marking list formed by marking a plurality of asset data to be marked to obtain marking labels, and each asset data to be marked and the marking label corresponding to each asset data to be marked. And acquiring a preset asset marking list through a preset data acquisition interface, and providing data preparation for recommending more accurate asset data later. Wherein the preset data acquisition interface may be a read () function.

As shown in fig. 3, before the preset asset marking list is acquired, the business data processing method further includes, but is not limited to, the following steps:

step S810, obtaining a plurality of asset data to be marked, and preprocessing each asset data to be marked to obtain a plurality of preprocessed asset data.

In an embodiment, the asset data to be marked includes different data assets such as products, markets, orders, costs, incomes, services, channels, etc., and a plurality of asset data to be marked can be obtained through a crawler technology, or can be obtained through sorting historical data of banking businesses. And then the asset data to be marked are arranged and stored in scv, txt or other table forms, the asset data to be marked is obtained by utilizing a data reading interface, and data support is provided for the asset marking list obtained by marking in the follow-up process.

In one embodiment, each asset data to be marked is pre-processed, including data cleansing and vectorization of the data. The data cleaning performs data screening on the plurality of asset data to be marked, removes disordered data and the like, and the data screening selects data columns, such as field Chinese names, field English names, table Chinese names, table English names, field types, field lengths and the like, of the asset data to be marked, so that important field data are screened out, and the calculated amount can be reduced under the condition of ensuring accuracy. The removal of the clutter data may be the removal of empty lines, clutter, remarks, duplicate data, etc. The repeated record row can be removed according to the fields such as category, field Chinese name, field English name and the like, and only one row of record is reserved in the same data. The data preprocessing is performed to obtain a plurality of preprocessed asset data, so that the calculated amount can be reduced, and the marking precision can be improved.

And step S820, marking each piece of preprocessed asset data by using the target marking model to obtain marking labels.

In an embodiment, the target marking model is a trained model, the target marking model is utilized to mark the preprocessed asset data obtained in the step S810, the preprocessed asset data is directly input into the target marking model, and the target marking model outputs marking labels, so that the asset marking list is conveniently obtained through subsequent construction.

Step S830, obtaining an asset marking list according to each asset data to be marked and the marking labels corresponding to each asset data to be marked.

In an embodiment, each asset data to be marked and the marking tag corresponding to the asset data to be marked are used as a record, a plurality of asset data to be marked and the marking tag corresponding to the asset data to be marked form a plurality of records, and the plurality of records form an asset marking list. Because the plurality of asset data to be marked are data with different attributes in the industry, the asset marking list comprises richer asset data and marking labels thereof, the asset marking list is favorable for matching with related recommendation results by subsequent utilization, and accurate asset data are recommended for users.

As shown in fig. 4, before marking each piece of preprocessed asset data by using the target marking model, the business data processing method further includes, but is not limited to, the following steps:

step S910, an asset sample data set is acquired, where the asset sample data set includes a plurality of asset sample data and marking tags corresponding to the respective asset sample data.

In one embodiment, the asset sample data set includes a plurality of asset sample data including different data assets of products, markets, orders, costs, revenues, services, channels, etc., which may be obtained by crawler technology and then consolidated to save the plurality of asset sample data in scv, txt or other tabular form. The asset sample data set further comprises marking labels corresponding to the asset sample data, wherein the marking labels are obtained by manually marking a plurality of asset sample data, and the marking labels are stored in a file corresponding to the asset sample data. And acquiring an asset sample data set by using a data reading interface, and providing data support for subsequent model training.

The method is characterized in that the asset sample data is manually marked, marking rules are divided according to standards published by banking industry, and the asset sample data is matched with corresponding subclasses to obtain marking labels. As shown in fig. 5, the categories include: the method comprises the steps of participant general purpose, personal general purpose, organization general purpose, customer information-basic information, risk information-credit risk, management personality-personal client, enterprise client, financial staff client, employee information, institution information, merchant information-basic information, terminal information and the like, wherein each category corresponds to a corresponding grade, for example, merchant information-basic information corresponds to grade 2, when asset sample data are input as merchant information, a subclass under the merchant information is further determined, the basic information subclass is further determined, and then a label marked by the asset sample data is expressed as merchant information-basic information-/-2.

Step S920, preprocessing is performed on each asset sample data to obtain a plurality of preprocessed sample data.

In one embodiment, each asset sample data is pre-processed, including data cleansing and vectorization of the data. The data cleaning performs data screening on a plurality of asset sample data, removes clutter data and the like, and the data screening screens important field data for selecting data columns, such as field Chinese names, field English names, table Chinese names, table English names, field types, field lengths and the like, of the asset sample data, which have influence on marking classification, so that the calculation amount can be reduced under the condition of ensuring accuracy. The removal of the clutter data may be the removal of empty lines, clutter, remarks, duplicate data, etc. The repeated record row can be removed according to the fields such as category, field Chinese name, field English name and the like, and only one row of record is reserved in the same data. The data preprocessing is performed to obtain a plurality of preprocessed sample data, so that the calculated amount can be reduced, and the accuracy of model marking is improved.

In step S930, a first sample equalization process is performed on each of the preprocessed sample data, to obtain a training data set.

As shown in fig. 6, the first sample equalization process is performed on each of the preprocessed sample data to obtain a training data set, including, but not limited to, the following steps:

step S931, performing de-duplication processing on the preprocessed sample data according to the category, the field Chinese name and the field English name to obtain first de-duplication sample data.

In some possible embodiments of the present application, the deduplication processing is performed according to the category, the field chinese name, and the field english name according to the preprocessed sample data obtained in step S920. For example, under the condition that the types, the field Chinese names and the field English names of the pieces of preprocessed sample data are the same, the pieces of preprocessed sample data are repeated data, the repeated data are deleted, only one piece of preprocessed sample data in the same data is ensured, and the first duplicate removal sample data is obtained, so that the subsequent training set according to the first duplicate removal sample data is facilitated.

Step S932, selecting N categories with the number being arranged in front from the first de-duplicated sample data, to obtain selected sample data, where N is an integer greater than or equal to 2.

In some possible embodiments of the present application, N is an integer greater than or equal to 2, and N is not greater than 10. And selecting N categories with the highest number from the first de-duplicated sample data, and selecting the 2 categories with the highest number when N is 2 to obtain selected sample data, so that secondary de-duplication is facilitated according to the selected sample data. The category is a major category, and a plurality of minor categories are included under the major category.

Step S933, the selected sample data is de-duplicated according to the category and the field Chinese name to obtain second de-duplicated sample data.

In some possible embodiments of the present application, since the field chinese name may be translated in different english, the duplicate removal is performed again according to the category and the field chinese name on the basis of the selected sample obtained in step S932, so as to obtain the second duplicate removal sample data. The presence of data having the same meaning in English can be avoided, and the repeated data can be formed, so that the second duplicate removal sample data is obtained without repeated data.

Step S934, removing the categories with the number not exceeding the preset number in the second duplicate removal sample data to obtain a training data set.

In some possible embodiments of the present application, according to step S933, the second de-duplicated sample data is obtained, the number of the second de-duplicated sample data is counted, in each sub-class, the classes with the number not exceeding the preset number in the second de-duplicated sample data are removed, the preset number may be 10 or 20, and the classes with the number less than or equal to 10 in the second de-duplicated sample data are removed, so as to obtain the training data set. The training data set is constructed through the steps, so that the accuracy of the model is improved.

Step S940, performing a second sample equalization process on each of the preprocessed sample data, to obtain a verification data set.

In one embodiment, the second sample equalization process is performed on each of the preprocessed sample data to obtain the verification data set, which specifically includes, but is not limited to: the categories with the number less than or equal to 10 in the preprocessed sample data are removed, the number of the categories does not exceed the preset number, the preset number can be 10 or 20, and the categories with the number less than or equal to 10 in the preprocessed sample data are removed, so that a verification data set is obtained, and the accuracy of the model is improved.

As shown in fig. 7, assuming that there are 5789 pieces of preprocessed sample data, performing duplication removal processing on the preprocessed sample data according to category, field chinese name and field english name, and selecting the first 2 categories from the duplicated data; performing duplication removal on the selected data according to the category and the field Chinese name to obtain 2252 data volumes after duplication removal; then, the categories with the number not exceeding 10 in the data are removed, and a training data set is obtained, wherein the training data set has 2250 pieces of data. The number of categories of no more than 10 are removed from the pre-processed sample data to obtain a validation data set having data 5780.

In step S950, the training data set is used to tune the preset first initial marking model to obtain a second initial marking model.

In an embodiment, the training data set includes preprocessed sample data which has no repetition and has a data amount meeting a preset number, and the marking label corresponding to the preprocessed sample data, and the preprocessed sample data and the marking label are used for performing parameter adjustment on a preset first initial marking model to obtain a second initial marking model, so that the target marking model can be obtained later. The first initial marking model can be a bert model, can be a variant of the bert model, can be other deep learning models, and can meet the accurate marking requirement, and is not described in detail herein; the model parameter adjustment is a parameter adjustment mode of the bert model, and is not described herein.

Step S960, verifying the second initial marking model by using the verification data set, and obtaining the target marking model under the condition that the accuracy is larger than a preset accuracy threshold.

In an embodiment, the training data set includes preprocessed sample data meeting a preset number, and the marking labels corresponding to the preprocessed sample data are not included, the verification data set is used for verifying the second initial marking model, and under the condition that the accuracy is greater than a preset accuracy threshold, the model is better in generalization, the target marking model is obtained, and the accuracy of marking the retrieved asset data is higher.

In another embodiment, under the condition that the accuracy is smaller than or equal to a preset accuracy threshold, performing data dimension-lifting processing on the training data set, and performing parameter adjustment on the first initial marking model by using the training data set after dimension lifting, so that the accuracy is larger than the preset accuracy threshold.

In yet another embodiment, when the bank provides new sample data, the sample data is no corresponding marking label, the categories of which the number is not more than the preset number in the new sample data are removed, and the target marking model is tested, so that the accuracy is up to 82%, and the target marking model has good generalization.

And S500, obtaining a matching result according to the recommended data, the label corresponding to the recommended data, the asset data to be marked and the marking label.

In an embodiment, the matching result is obtained according to the recommended data, the tag corresponding to the recommended data, the asset data to be marked and the marking tag, which specifically includes but is not limited to: firstly, classifying tags corresponding to recommended data and marking tags, judging whether the tags corresponding to the recommended data and the marking tags are of the same subclass by using a preset classification algorithm, performing similarity calculation on the recommended data and the asset data to be marked by using a similarity algorithm, screening out data with the consistent tag class and similarity smaller than a preset threshold value, and obtaining a matching result, thereby being beneficial to follow-up accurate asset data recommendation according to the matching result. The classification algorithm can be a support vector machine, a deep learning algorithm and the like; the similarity algorithm may be a euclidean distance, a character similarity algorithm, and the like, and will not be described herein.

And step S600, performing second recommendation according to the matching result and the retrieved asset data to obtain recommended asset data.

In an embodiment, the second recommendation is an accurate recommendation performed according to the matching result, and in the data of the same category, the asset data more meeting the user requirement is recommended. The matching result obtained in the step S500 is favorable to be that the similarity is high, the labels are the same, the matching result is highly matched with the retrieval asset data, the first recommended data are recommended and arranged, and the recommended asset data are obtained. The recommended asset data is screened for multiple times, so that the related requirements of the retrieved asset data are met, and accurate asset data are recommended for users.

The corresponding tag in the recommended asset data can be used as the tag of the retrieved asset data, the retrieved asset data and the corresponding tag are added into the asset marking list, the data amount in the asset marking list is increased, and the recommendation accuracy can be improved.

As shown in fig. 8, fig. 8 shows an overall flow diagram of a business data processing method provided in the embodiment of the present application, first, search asset data is obtained, vector processing is performed on the search asset data in response to a selection operation on a preset target marking model, dimension raising processing is performed on the search asset data, and the processed data are fused and input into the target marking model for marking processing, so as to obtain a marking result, and the obtained marking result indicates a category to which the marking result belongs, thereby being beneficial to accurate recommendation; according to the marking result, performing first recommendation by using one strategy of preset quantity limitation, matching degree limitation and combination limitation to obtain a related recommendation result, wherein the related recommendation result comprises recommendation data and labels corresponding to the recommendation data, and performing rough recommendation by using the recommendation strategy, so that recommendation accuracy is improved.

In an embodiment, acquiring a plurality of asset data to be marked, preprocessing each asset data to be marked to obtain a plurality of preprocessed asset data, and marking each preprocessed asset data by using a target marking model to obtain marking labels; and confirming and publishing the marked data, checking the asset data, realizing effective management of the assets, and obtaining an asset marking list according to the confirmed and published asset data to be marked and marking labels corresponding to the asset data to be marked. Acquiring a preset asset marking list, wherein the asset marking list comprises a plurality of asset data to be marked and marking labels corresponding to the asset data to be marked; obtaining a matching result according to the recommended data, the label corresponding to the recommended data, the asset data to be marked and the marking label; and performing second recommendation according to the matching result and the retrieved asset data to obtain recommended asset data, and performing the second recommendation can obtain more accurate recommended asset data.

As shown in fig. 9, an embodiment of the present application provides a service data processing apparatus 100, where the service data processing apparatus 100 obtains, through a first data obtaining module 110, search asset data, where the search asset data is search content input by a user, and obtains the search asset data to provide data support for recommending accurate asset data to the user later; the marking module 120 is utilized to respond to the selection operation of the preset target marking model, the target marking model is utilized to mark the retrieved asset data to obtain a marking result, the retrieved asset data is firstly marked, and the obtained marking result indicates the category to which the retrieved asset data belongs, so that accurate recommendation is facilitated; the first recommendation module 130 is adopted to conduct first recommendation according to the marking result and a preset recommendation strategy, so that a related recommendation result is obtained, the related recommendation result comprises recommendation data and labels corresponding to the recommendation data, and coarse recommendation is conducted by utilizing the recommendation strategy, so that recommendation accuracy is improved; acquiring a preset asset marking list by adopting a second data acquisition module 140, wherein the asset marking list comprises a plurality of asset data to be marked and marking labels corresponding to the asset data to be marked, and providing data support for follow-up accurate recommendation; then, a matching module 150 is utilized to obtain a matching result according to the recommended data, the label corresponding to the recommended data, the asset data to be marked and the marking label; and finally, performing second recommendation by using the second recommendation module 160 according to the matching result and the retrieved asset data to obtain recommended asset data, so that more accurate recommended asset data can be obtained.

It should be noted that, the first data obtaining module 110 is connected to the marking module 120, the marking module 120 is connected to the first recommending module 130, the first recommending module 130 is connected to the second data obtaining module 140, the second data obtaining module 140 is connected to the matching module 150, and the matching module 150 is connected to the second recommending module 160. The business data processing method is applied to the business data processing device 100, the business data processing device 100 performs marking on the retrieved asset data through the target marking model, performs first recommendation according to the marking result, screens out asset data similar to the retrieved asset data, performs second recommendation according to the asset marking list, and can recommend accurate asset data for a user through multi-layer recommendation.

Also to be described is: in the device provided in the above embodiment, when implementing the functions thereof, only the division of the above functional modules is used as an example, in practical application, the above functional allocation may be implemented by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules, so as to implement all or part of the functions described above. In addition, the embodiments of the apparatus and the method provided in the foregoing embodiments belong to the same concept, and specific implementation processes of the embodiments of the method are detailed in the method embodiments, which are not repeated herein.

The application also discloses electronic equipment. Referring to fig. 10, fig. 10 is a schematic structural view of an electronic device according to the disclosure of the embodiment of the present application. The electronic device 500 may include: at least one processor 501, at least one network interface 504, a user interface 503, a memory 505, at least one communication bus 502.

Wherein a communication bus 502 is used to enable connected communications between these components.

The user interface 503 may include a Display screen (Display) and a Camera (Camera), and the optional user interface 503 may further include a standard wired interface and a standard wireless interface.

The network interface 504 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface), among others.

Wherein the processor 501 may include one or more processing cores. The processor 501 connects various parts throughout the server using various interfaces and lines, performs various functions of the server and processes data by executing or executing instructions, programs, code sets, or instruction sets stored in the memory 505, and invoking data stored in the memory 505. Alternatively, the processor 501 may be implemented in hardware in at least one of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), programmable logic array (Programmable Logic Array, PLA). The processor 501 may integrate one or a combination of several of a central processing unit (Central Processing Unit, CPU), an image processor (Graphics Processing Unit, GPU), and a modem, etc. The CPU mainly processes an operating system, a user interface, an application program and the like; the GPU is used for rendering and drawing the content required to be displayed by the display screen; the modem is used to handle wireless communications. It will be appreciated that the modem may not be integrated into the processor 501 and may be implemented by a single chip.

The Memory 505 may include a random access Memory (Random Access Memory, RAM) or a Read-Only Memory (Read-Only Memory). Optionally, the memory 505 comprises a non-transitory computer readable medium (non-transitory computer-readable storage medium). Memory 505 may be used to store instructions, programs, code sets, or instruction sets. The memory 505 may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing the above-described various method embodiments, etc.; the storage data area may store data or the like involved in the above respective method embodiments. The memory 505 may also optionally be at least one storage device located remotely from the processor 501. Referring to fig. 10, an operating system, a network communication module, a user interface module, and an application program of a service data processing method may be included in the memory 505 as a computer storage medium.

In the electronic device 500 shown in fig. 10, the user interface 503 is mainly used for providing an input interface for a user, and acquiring data input by the user; and the processor 501 may be configured to invoke an application program in the memory 505 that stores a business data processing method, which when executed by the one or more processors 501, causes the electronic device 500 to perform the method as in one or more of the embodiments described above. It should be noted that, for simplicity of description, the foregoing method embodiments are all expressed as a series of action combinations, but it should be understood by those skilled in the art that the present application is not limited by the order of actions described, as some steps may be performed in other order or simultaneously in accordance with the present application. Further, those skilled in the art will also appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily required in the present application.

In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to related descriptions of other embodiments.

In the several embodiments provided herein, it should be understood that the disclosed apparatus may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, such as a division of units, merely a division of logic functions, and there may be additional divisions in actual implementation, such as multiple units or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some service interface, device or unit indirect coupling or communication connection, electrical or otherwise.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable memory. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a memory, including several instructions for causing a computer device (which may be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the methods of the embodiments of the present application. And the aforementioned memory includes: various media capable of storing program codes, such as a U disk, a mobile hard disk, a magnetic disk or an optical disk.

The above are merely exemplary embodiments of the present disclosure and are not intended to limit the scope of the present disclosure. That is, equivalent changes and modifications are contemplated by the teachings of this disclosure, which fall within the scope of the present disclosure. Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure.

This application is intended to cover any adaptations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a scope and spirit of the disclosure being indicated by the claims.

Claims

1. A method for processing service data, the method comprising:

acquiring retrieval asset data;

2. The method of claim 1, wherein prior to the obtaining the pre-set asset tag list, the method further comprises:

3. The method of claim 2, wherein prior to said marking each of said preprocessed asset data with said target marking model to obtain a marked tag, said method further comprises:

4. A method according to claim 3, wherein said performing a first sample equalization process on each of said preprocessed sample data results in a training data set, comprising:

5. The method of claim 1, wherein prior to said marking said retrieved asset data with said target marking model to obtain a marking result, said method further comprises:

6. The method of claim 1, wherein the obtaining a matching result according to the recommendation data, the tag corresponding to the recommendation data, the asset data to be marked, and the marking tag comprises:

7. The method of claim 1, wherein the recommended policy is at least one of a quantity limit, a matching degree limit, a combination limit.

8. A traffic data processing apparatus, the apparatus comprising:

a first data acquisition module (110) for acquiring search asset data;

the marking module (120) is used for marking the retrieved asset data by utilizing the target marking model in response to the selection operation of the preset target marking model to obtain a marking result;

the first recommendation module (130) is used for carrying out first recommendation by utilizing a preset recommendation strategy according to the marking result to obtain a related recommendation result, wherein the related recommendation result comprises recommendation data and labels corresponding to the recommendation data;

the second data acquisition module (140) is used for acquiring a preset asset marking list, wherein the asset marking list comprises a plurality of asset data to be marked and marking labels corresponding to the asset data to be marked;

the matching module (150) is used for obtaining a matching result according to the recommended data, the label corresponding to the recommended data, the asset data to be marked and the marking label;

And the second recommendation module (160) is used for carrying out second recommendation according to the matching result and the retrieval asset data to obtain recommended asset data.

9. An electronic device comprising a processor (501), a memory (505), a user interface (503), a communication bus (502) and a network interface (504), the processor (501), the memory (505), the user interface (503) and the network interface (504) being respectively connected to the communication bus (502), the memory (505) being adapted to store instructions, the user interface (503) and the network interface (504) being adapted to communicate to other devices, the processor (501) being adapted to execute the instructions stored in the memory (505) to cause the electronic device (500) to perform the method according to any of claims 1-7.

10. A computer readable storage medium storing instructions which, when executed, perform the method of any one of claims 1-7.