A service problem attribution method and device

Abstract

The embodiment of the present application provides a service problem attribution method and device, which adopts a service problem attribution model to analyze the service problem classification results corresponding to the text data to be analyzed. The service problem attribution model analyzes and obtains the service problem classification results for the text data to be analyzed based on the text data to be analyzed and the service knowledge graph related to the text data to be analyzed. The service problem attribution model provided in the above method takes the text data to be analyzed and the service knowledge graph related to the text data to be analyzed as input information. The service knowledge graph can obtain the attribute information of the service information in the text data to be analyzed, thereby improving the understanding of the language expression logic of the text data to be analyzed. Based on the service knowledge graph, it helps to improve the accuracy of the service problem attribution model in determining the language expression logic of the text data to be analyzed, and the accuracy of the service problem attribution model in matching the service problem classification results to the text data to be analyzed is also improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and specifically to a service problem attribution method. The present application also relates to a service problem attribution device, an electronic device and a computer storage medium, another service problem attribution method, device, electronic device and computer storage medium, and a training method, device, electronic device and computer storage medium for a service problem attribution model. Background Art

[0002] Currently, merchants provide various services to users on online platforms. After using the services, users evaluate the services on the online platforms. In order to quickly analyze the problems in the evaluation information provided by users for service information, the online platform uses an analysis model to analyze the evaluation information provided by users and identify the problem types corresponding to the evaluation information.

[0003] In the prior art, the above-mentioned method of analyzing the types of questions in merchant service information usually has the problem of incorrect recognition results and low recognition accuracy. Therefore, how to improve the accuracy of the types of questions in the analysis and evaluation information is a problem that needs to be solved. Summary of the invention

[0004] The embodiment of the present application provides a service problem attribution method to improve the accuracy of problem types in analysis and evaluation information. The present application also relates to a service problem attribution device, electronic device and computer storage medium, another service problem attribution method, device, electronic device and computer storage medium, and a service problem attribution model training method, device, electronic device and computer storage medium.

[0005] An embodiment of the present application provides a service problem attribution method, comprising: obtaining text data to be analyzed for service information; inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed; wherein the service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.

[0006] Optionally, the step of inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model to obtain a first embedding vector for the text data to be analyzed, the first embedding vector including an embedding vector of the text data to be analyzed and an embedding vector of a service knowledge graph related to the text data to be analyzed; and obtaining a service problem classification result corresponding to the text data to be analyzed according to the first embedding vector.

[0007] Optionally, the step of inputting the text data to be analyzed into the service problem attribution model to obtain a first embedding vector for the text data to be analyzed includes: inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector for the text data to be analyzed; querying a service knowledge graph embedding vector associated with the embedding vector for the text data to be analyzed based on the embedding vector for the text data to be analyzed; and obtaining a first embedding vector for the text data to be analyzed based on the embedding vector for the text data to be analyzed and the embedding vector for the service knowledge graph.

[0008] Optionally, obtaining a service problem classification result corresponding to the text data to be analyzed based on the first embedding vector includes: obtaining feature information of the text data to be analyzed based on the first embedding vector; and searching for a target service problem classification result that matches the feature information of the text data to be analyzed in a service problem classification list of the service problem attribution model based on the feature information of the text data to be analyzed, as the service problem classification result corresponding to the text data to be analyzed.

[0009] Optionally, according to the feature information of the text data to be analyzed, searching the service problem classification list of the service problem attribution model for a target service problem classification result that matches the feature information of the text data to be analyzed as the service problem classification result corresponding to the text data to be analyzed, including: obtaining feature information corresponding to multiple candidate service problem classification results in the service problem classification list of the service problem attribution model; comparing the feature information of the text data to be analyzed with the feature information corresponding to the multiple candidate service problem classification results, and determining the candidate service problem classification result containing the feature information of the text data to be analyzed as the target service problem classification result that matches the feature information of the text data to be analyzed as the service problem classification result corresponding to the text data to be analyzed.

[0010] Optionally, obtaining a service problem classification result corresponding to the text data to be analyzed based on the first embedding vector includes: obtaining feature information of the text data to be analyzed based on the first embedding vector; searching for at least one service problem classification result that matches the feature information of the text data to be analyzed in a service problem classification list of the service problem attribution model based on the feature information of the text data to be analyzed; and using the at least one service problem classification result as the service problem classification result corresponding to the text data to be analyzed.

[0011] Optionally, obtaining a service problem classification result corresponding to the text data to be analyzed according to the first embedding vector includes: obtaining feature information of the text data to be analyzed according to the first embedding vector; querying a service problem major classification result associated with the feature information of the text data to be analyzed in a service problem major classification list in the service problem attribution classification model according to the feature information of the text data to be analyzed as a target service problem major classification result of the text data to be analyzed; querying a target service problem minor classification result associated with the feature information of the text data to be analyzed in a service problem minor classification list of the target service problem major classification result on the basis of determining the target service problem major classification result of the text data to be analyzed; and generating a service problem classification result corresponding to the text data to be analyzed according to the target service problem major classification result and the target service problem minor classification result.

[0012] Optionally, the service problem attribution model includes a major classification text data feature analysis model, which is used to analyze feature information of text data corresponding to service problem major classification results; according to the feature information of the text data to be analyzed, querying the service problem major classification results associated with the feature information of the text data to be analyzed in the service problem major classification list in the service problem attribution classification model as the target service problem major classification results of the text data to be analyzed, including: according to the major classification text data feature analysis model, obtaining first feature information of the text data used to analyze the service problem major classification results in the text data to be analyzed; obtaining first candidate feature information of the text data corresponding to each candidate service problem major classification result in the service problem major classification list; comparing the first feature information in the text data to be analyzed with the first candidate feature information corresponding to each candidate service problem major classification result, to obtain the target service problem major classification result corresponding to the text data to be analyzed.

[0013] Optionally, based on determining the target service problem major classification result of the text data to be analyzed, searching the service problem minor classification list of the target service problem major classification result for the target service problem minor classification result associated with the feature information of the text data to be analyzed, including: obtaining second feature information of the text data used to analyze the service problem minor classification result in the text data to be analyzed; obtaining second candidate feature information of the text data corresponding to each candidate service problem minor classification result in the service problem minor classification list; comparing the second feature information in the text data to be analyzed with the second candidate feature information corresponding to each candidate service problem minor classification result to obtain the target service problem minor classification result corresponding to the text data to be analyzed.

[0014] Optionally, the step of inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector of the text data to be analyzed includes: inputting the text data to be analyzed into the service problem attribution model to obtain a text embedding vector corresponding to each text unit in the text data to be analyzed, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a position embedding vector corresponding to each text unit in the text data to be analyzed; processing the text embedding vector corresponding to each text unit in the text data to be analyzed, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed to obtain the embedding vector of the text data to be analyzed.

[0015] Optionally, querying the service knowledge graph embedding vector associated with the text data embedding vector to be analyzed based on the text data embedding vector to be analyzed includes: obtaining the service information embedding vector in the text data to be analyzed based on the text data embedding vector to be analyzed; querying the service knowledge graph for the service information based on the service information embedding vector in the text data to be analyzed; obtaining the text embedding vector corresponding to each text unit in the service knowledge graph for the service information, the paragraph embedding vector corresponding to each text unit in the service knowledge graph, and the position embedding vector corresponding to each text unit in the service knowledge graph; processing the text embedding vector corresponding to each text unit in the service knowledge graph for the service information, the paragraph embedding vector corresponding to each text unit in the service knowledge graph, and the position embedding vector corresponding to each text unit in the service knowledge graph to obtain the service knowledge graph embedding vector for the service information.

[0016] Optionally, the service information is food service information for food services provided by the merchant to the user; the analysis text data for the service information is the user's evaluation information for the food service information; and obtaining the text data to be analyzed for the service information includes: obtaining the text data to be analyzed for the food service information sent by the user end.

[0017] Optionally, the service problem attribution model is specifically used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed, the pinyin information corresponding to the text data to be analyzed, and the service knowledge graph related to the text data to be analyzed.

[0018] Optionally, the step of inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model to obtain a second embedding vector for the text data to be analyzed, the second embedding vector including a pinyin embedding vector of the text data to be analyzed and a pinyin embedding vector of a service knowledge graph related to the text data to be analyzed; and obtaining a service problem classification result corresponding to the text data to be analyzed according to the second embedding vector.

[0019] Optionally, the step of inputting the text data to be analyzed into the service problem attribution model to obtain a second embedding vector for the text data to be analyzed comprises: inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector for the text data to be analyzed, the embedding vector for the text data to be analyzed including a pinyin embedding vector corresponding to each text unit in the text data to be analyzed; according to the embedding vector for the text data to be analyzed, querying a service knowledge graph embedding vector associated with the embedding vector for the text data to be analyzed, the service knowledge graph embedding vector including a pinyin embedding vector corresponding to each text unit in the service knowledge graph; and obtaining a second embedding vector for the text data to be analyzed according to the embedding vector for the text data to be analyzed and the service knowledge graph embedding vector.

[0020] Optionally, the step of inputting the text data to be analyzed into the service problem attribution model to obtain the embedding vector of the text data to be analyzed includes: inputting the text data to be analyzed into the service problem attribution model to obtain the text embedding vector corresponding to each text unit in the text data to be analyzed, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed; processing the text embedding vector corresponding to each text unit in the text data to be analyzed, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed to obtain the embedding vector of the text data to be analyzed.

[0021] Optionally, querying the service knowledge graph embedding vector associated with the text data embedding vector to be analyzed based on the text data embedding vector to be analyzed includes: obtaining the service information embedding vector in the text data to be analyzed based on the text data embedding vector to be analyzed; querying the service knowledge graph for the service information based on the service information embedding vector in the text data to be analyzed; obtaining the text embedding vector corresponding to each text unit in the service knowledge graph for the service information, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the service knowledge graph, and the position embedding vector corresponding to each text unit in the service knowledge graph; processing the text embedding vector corresponding to each text unit in the service knowledge graph for the service information, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the service knowledge graph, and the position embedding vector corresponding to each text unit in the service knowledge graph to obtain the service knowledge graph embedding vector for the service information.

[0022] An embodiment of the present application also provides a service problem attribution method, comprising: obtaining text data to be analyzed for service information; inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed; wherein the service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and pinyin information corresponding to the text data to be analyzed.

[0023] Optionally, the step of inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model to obtain an embedding vector of the text data to be analyzed, the embedding vector of the text data to be analyzed including a text embedding vector corresponding to each text unit in the text data to be analyzed, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a position embedding vector corresponding to each text unit in the text data to be analyzed; and obtaining a service problem classification result corresponding to the text data to be analyzed according to the embedding vector of the text data to be analyzed.

[0024] Optionally, the step of inputting the text data to be analyzed into the service problem attribution model to obtain the embedding vector of the text data to be analyzed includes: inputting the text data to be analyzed into the service problem attribution model, performing vectorization processing on the text data to be analyzed, and obtaining a text embedding vector corresponding to each text unit in the text data to be analyzed, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a position embedding vector corresponding to each text unit in the text data to be analyzed; processing the text embedding vector corresponding to each text unit in the text data to be analyzed, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed to obtain the embedding vector of the text data to be analyzed.

[0025] Optionally, obtaining a service problem classification result corresponding to the text data to be analyzed based on the embedding vector of the text data to be analyzed includes: obtaining feature information of the text data to be analyzed based on the embedding vector of the text data to be analyzed; and searching for a target service problem classification result that matches the feature information of the text data to be analyzed in a service problem classification list of the service problem attribution model based on the feature information of the text data to be analyzed, as the service problem classification result corresponding to the text data to be analyzed.

[0026] Optionally, obtaining a service problem classification result corresponding to the text data to be analyzed based on the embedding vector of the text data to be analyzed includes: obtaining feature information of the text data to be analyzed based on the embedding vector of the text data to be analyzed; querying a service problem major classification result associated with the feature information of the text data to be analyzed in a service problem major classification list in the service problem attribution classification model based on the feature information of the text data to be analyzed, as a target service problem major classification result of the text data to be analyzed; querying a target service problem minor classification result associated with the feature information of the text data to be analyzed in a service problem minor classification list of the target service problem major classification result on the basis of determining the target service problem major classification result of the text data to be analyzed; generating a service problem classification result corresponding to the text data to be analyzed based on the target service problem major classification result and the target service problem minor classification result.

[0027] Optionally, the service problem attribution model is specifically used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed, the pinyin information corresponding to the text data to be analyzed, and the service knowledge graph related to the text data to be analyzed.

[0028] Optionally, the step of inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model to obtain a second embedding vector for the text data to be analyzed, the second embedding vector including an embedding vector of the text data to be analyzed and an embedding vector of a service knowledge graph related to the text data to be analyzed; and obtaining a service problem classification result corresponding to the text data to be analyzed according to the second embedding vector.

[0029] Optionally, the step of inputting the text data to be analyzed into the service problem attribution model to obtain a second embedding vector for the text data to be analyzed comprises: inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector for the text data to be analyzed, the embedding vector for the text data to be analyzed including a pinyin embedding vector corresponding to each text unit in the text data to be analyzed; according to the embedding vector for the text data to be analyzed, querying a service knowledge graph embedding vector associated with the embedding vector for the text data to be analyzed, the service knowledge graph embedding vector including a pinyin embedding vector corresponding to each text unit in the service knowledge graph; and obtaining a second embedding vector for the text data to be analyzed according to the embedding vector for the text data to be analyzed and the service knowledge graph embedding vector.

[0030] Optionally, obtaining the service problem classification result corresponding to the text data to be analyzed according to the second embedding vector includes: obtaining feature information of the text data to be analyzed according to the second embedding vector; querying the service problem major classification result associated with the feature information of the text data to be analyzed in the service problem major classification list in the service problem attribution classification model according to the feature information of the text data to be analyzed as a target service problem major classification result of the text data to be analyzed; querying the target service problem minor classification result associated with the feature information of the text data to be analyzed in the service problem minor classification list of the target service problem major classification result on the basis of determining the target service problem major classification result of the text data to be analyzed; generating the service problem classification result corresponding to the text data to be analyzed according to the target service problem major classification result and the target service problem minor classification result.

[0031] The embodiment of the present application also provides a training method for a service problem attribution model, including: obtaining a pre-trained model for analyzing feature information of text data; adjusting the pre-trained model according to text data samples for service information and service problem major classification result samples for the text data samples to obtain a major classification text data feature analysis model, wherein the major classification text data feature analysis model is used to analyze feature information of text data corresponding to the major classification result of service problems; adjusting the major classification text data feature analysis model according to text data samples for service information and service problem minor classification result samples for the text data samples to obtain a service problem attribution model for analyzing service problem classification results for text data to be analyzed.

[0032] Optionally, the pre-trained model analyzes feature information of text data in the following manner: obtaining a text data embedding vector based on the text data, the text data embedding vector including a text embedding vector corresponding to each text unit in the text data, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector of each text unit in the text data, and a position embedding vector of each text unit in the text data; analyzing the feature information of the text data based on the text data embedding vector.

[0033] Optionally, it also includes: obtaining a service knowledge graph embedding vector associated with the text data based on the text data embedding vector, the service knowledge graph embedding vector including a text embedding vector corresponding to each text unit in the service knowledge graph associated with the text data, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector of each text unit in the service knowledge graph, and a position embedding vector of each text unit in the service knowledge graph; analyzing feature information of the text data based on the text data embedding vector includes: analyzing feature information of the text data based on the text data embedding vector and the service knowledge graph embedding vector associated with the text data.

[0034] Optionally, the pre-trained model is adjusted according to the text data sample for service information and the service problem classification result sample for the text data sample to obtain the large classification text data feature analysis model, including: inputting the text data sample for service information into the pre-trained model to obtain a first service problem classification result for the text data sample for the service information output by the pre-trained model; and adjusting the service problem classification parameters of the pre-trained model according to the similarity between the first service problem classification result and the service problem classification result sample for the text data sample to obtain the large classification text data feature analysis model.

[0035] Optionally, the step of inputting the text data sample for the service information into the pre-trained model to obtain a first major classification result of service problems for the text data sample for the service information output by the pre-trained model comprises: inputting the text data sample for the service information into the pre-trained model to obtain feature information of the text data sample; and obtaining a first major classification result of service problems for the text data sample for the service information output by the pre-trained model based on the feature information of the text data sample.

[0036] Optionally, the large classification text data feature analysis model is adjusted according to the text data sample for service information and the service problem sub-classification result sample for the text data sample to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed, including: inputting the text data for the service information into the large classification text data feature analysis model to obtain a second service problem sub-classification result for the text data sample output by the large classification text data feature analysis model; and adjusting the large classification text data feature analysis model according to the degree of similarity between the second service problem sub-classification result and the service problem sub-classification result sample to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed.

[0037] Optionally, it also includes: taking text data samples for service information and samples of large classification results of service problems that do not belong to the text data as a first negative sample pair, adjusting the pre-trained model, and obtaining a large classification text data feature analysis model for analyzing the feature information of text data corresponding to the large classification results of service problems; taking text data samples for service information and samples of small classification results of service problems that do not belong to the text data as a second negative sample pair, adjusting the large classification text data feature analysis model, and obtaining a service problem attribution model for analyzing the service problem classification results of the text data to be analyzed.

[0038] An embodiment of the present application also provides a service problem attribution device, comprising: a first obtaining unit, used to obtain text data to be analyzed for service information; a second obtaining unit, used to input the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed; wherein the service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.

[0039] An embodiment of the present application also provides a service problem attribution device, including: a third obtaining unit, used to obtain text data to be analyzed for service information; a fourth obtaining unit, used to input the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed; wherein the service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and the pinyin information corresponding to the text data to be analyzed.

[0040] The embodiment of the present application also provides a training device for a service problem attribution model, including: a pre-trained model acquisition unit, used to obtain a pre-trained model for analyzing feature information of text data; a large classification text data feature analysis model acquisition unit, used to adjust the pre-trained model according to text data samples for service information and service problem large classification result samples for the text data samples, to obtain a large classification text data feature analysis model, wherein the large classification text data feature analysis model is used to analyze feature information of text data corresponding to the service problem large classification result; a service problem attribution model acquisition unit, used to adjust the large classification text data feature analysis model according to text data samples for service information and service problem small classification result samples for the text data samples, to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed.

[0041] An embodiment of the present application further provides an electronic device, comprising a processor and a memory; a computer program is stored in the memory, and the processor executes the above method after running the computer program.

[0042] An embodiment of the present application further provides a computer storage medium, wherein the computer storage medium stores a computer program, and the computer program performs the above method when executed.

[0043] Compared with the prior art, the embodiments of the present application have the following advantages:

[0044] An embodiment of the present application provides a service problem attribution method, comprising: obtaining text data to be analyzed for service information; inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed; wherein the service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.

[0045] In one embodiment of the present application, the service problem attribution model analyzes the text data to be analyzed and the service knowledge graph related to the text data to be analyzed to obtain the service problem classification result for the text data to be analyzed. The service problem attribution model provided in the above method takes the text data to be analyzed and the service knowledge graph related to the text data to be analyzed as input information. The service knowledge graph can obtain the attribute information of the service information in the text data to be analyzed, thereby improving the understanding of the language expression logic of the text data to be analyzed. Based on the service knowledge graph, it is helpful to improve the accuracy of the service problem attribution model in determining the language expression logic of the text data to be analyzed, and the accuracy of the service problem attribution model in matching the text data to be analyzed with the service problem classification result is also improved.

[0046] In one embodiment of the present application, the service problem attribution model analyzes the text data to be analyzed and the pinyin information corresponding to the text data to be analyzed to obtain the service problem classification result for the text data to be analyzed. The service problem attribution model provided in the above method, when processing the text data to be analyzed, combines the pinyin information of the text data to be analyzed to improve the accuracy of the language expression logic of the text data to be analyzed. Based on the text data to be analyzed, it helps to improve the accuracy of the language expression logic of the text data to be analyzed, and the accuracy of the service problem attribution model matching the text data to be analyzed with the service problem classification result is also improved.

[0047] In one embodiment of the present application, a method for training a service problem attribution model is provided, comprising: obtaining a pre-trained model for analyzing feature information of text data; adjusting the pre-trained model according to text data samples for service information and service problem major classification result samples for the text data samples to obtain a major classification text data feature analysis model, wherein the major classification text data feature analysis model is used to analyze feature information of text data corresponding to the service problem major classification result; adjusting the major classification text data feature analysis model according to text data samples for service information and service problem minor classification result samples for the text data samples to obtain a service problem attribution model for analyzing the service problem classification results for the text data to be analyzed.

[0048] The above method obtains a pre-trained model. First, the pre-trained model is adjusted using text data samples and service problem large classification result samples for the text data samples to be analyzed to obtain a large classification text data feature analysis model. Then, according to the text data samples and the service problem small classification result samples for the text data samples, the large classification text data feature analysis model is adjusted to obtain a service problem attribution model for analyzing the service problem classification results for the text data to be analyzed. It can be seen that when the trained service problem attribution model obtains the service problem classification results of the text data to be analyzed, the service problem large classification results of the text data to be analyzed are first obtained. On the basis of determining the service problem large classification results of the text data to be analyzed, the service problem small classification results of the text data to be analyzed are determined from the multiple service problem small classification results corresponding to the target service problem large classification results. Therefore, the service problem classification results of the text data to be analyzed finally obtained include the service problem large classification results and the service problem small classification results. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is an application scenario diagram of the service problem attribution method provided in an embodiment of the present application.

[0050] Figure 2 A schematic diagram of the service problem classification results for text information in the service problem attribution model provided in an embodiment of the present application.

[0051] Figure 3 This is the display page of the physical examination center on the merchant side provided in the embodiment of the present application.

[0052] Figure 4 Detailed food safety risk information for merchant stores provided in the embodiments of this application.

[0053] Figure 5 for Figure 4 Data information on detailed food safety risk information.

[0054] Figure 6 for Figure 5 The level of food safety risk issues occurring in the merchants and the corresponding handling methods for the food safety risk levels.

[0055] Figure 7 A food safety risk learning method developed for merchants provided in an embodiment of the present application.

[0056] Figure 8 This is a flowchart of a service problem attribution method provided in the first embodiment of the present application.

[0057] Fig. 9 This is a schematic diagram of a service problem attribution device provided in the second embodiment of the present application.

[0058] Fig.10 This is a schematic diagram of another service problem attribution method provided in the third embodiment of the present application.

[0059] Fig.11 This is a schematic diagram of another service problem attribution device provided in the fourth embodiment of the present application.

[0060] Fig.12 This is a schematic diagram of a training method for a service problem attribution model provided in the fifth embodiment of the present application.

[0061] Fig.13 This is a schematic diagram of a training device for a service problem attribution model provided in the sixth embodiment of the present application.

[0062] Fig.14 This is a schematic diagram of an electronic device provided in the seventh embodiment of the present application. DETAILED DESCRIPTION

[0063] Many specific details are described in the following description to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present application, so the present application is not limited by the specific implementation disclosed below.

[0064] The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The descriptions used in this application and the appended claims, such as "a", "first", and "second", are not limitations on quantity or sequence, but are used to distinguish the same type of information from each other.

[0065] The embodiment of the present application first provides a service problem attribution method, device, electronic device and computer storage medium. The embodiment of the present application also provides another service problem attribution method, device, electronic device and computer storage medium, a training method, device, electronic device and computer storage medium for a service problem attribution model.

[0066] In order to more clearly demonstrate the service problem attribution method provided by this application, the application scenario of the service problem attribution method provided by this application is first introduced.

[0067] The service problem attribution method provided in this application can be applied to food safety problem classification scenarios. For example, an online to offline catering service system provides users with meal ordering services. Figure 1As an example, it is an application scenario diagram of the service problem attribution method provided in the embodiment of the present application. Users order meals and cancel meals through the user terminal 101, and users feedback relevant data on meals, such as evaluation information for canceled meals, food safety insurance claim evaluation information, and evaluation data for consumed meals, etc. Among them, the evaluation information includes user complaint information regarding safety issues of some foods. The service terminal 102 receives the evaluation information sent by the user terminal for the meal, and based on the evaluation information, determines whether the evaluation information is complaint information regarding food safety issues of the meal. If so, determines the type of food safety issue corresponding to the evaluation information, and sends the evaluation information and the type of food safety issue for the evaluation information to the merchant providing the meal.

[0068] The server 102 analyzes the food safety problem type corresponding to the evaluation information provided by the user through the service problem attribution model. Specifically, the obtained text data to be analyzed for the service information is input into the service problem attribution model to obtain the service problem classification result for the text data to be analyzed output by the service problem attribution model.

[0069] For example, the user's evaluation information on the meal is input into the service problem attribution model, and the service problem attribution model analyzes the evaluation information to determine the service problem classification result corresponding to the evaluation information, that is, the food safety attribution type corresponding to the meal. For example, the user's evaluation information on the meal is: "There are mold spots on the fruit in this fruit drink", and the evaluation information is input into the service problem attribution model. Then the service problem attribution model outputs the food safety attribution type corresponding to the meal as visible mold and deterioration. For another example, the evaluation information is: "There are cockroaches in this meal", and the evaluation information is input into the service problem attribution model. Then the service problem attribution model outputs the food safety attribution type corresponding to the meal as cockroaches.

[0070] The service problem attribution model is used to determine the service problem classification result corresponding to the evaluation information provided by the user for the service information. Specifically, the major classification result of the service problem corresponding to the evaluation information is first determined, and then the minor classification result of the service problem corresponding to the evaluation information is analyzed on the basis of the major classification result of the service problem. Based on the major classification result of the service problem and the minor classification result of the service problem in the major classification result of the service problem, the service problem classification result for the evaluation information is generated. Taking the above-mentioned evaluation information "There are cockroaches in this meal" as an example, the major classification result of the service problem of the evaluation information is foreign matter, and the minor classification result of the service problem is cockroaches, among which cockroaches are the minor classification result of foreign matter among pests and foreign matter.

[0071] After the server obtains the food safety risk classification results corresponding to the user's evaluation information, it sends the food safety incidents existing in the food provided by the merchant and the problem classification results corresponding to the food safety incidents to the merchant. The merchant checks its existing food safety risk issues through the physical examination center of the merchant.

[0072] like Figure 3 As shown, it is a display page of the merchant-side physical examination center provided in the embodiment of the present application. Figure 3 In the page shown, the physical examination center on the merchant side displays the merchant's risk diagnosis, management and control processing records, the merchant's food safety credit score information, and learning area information for food safety issues. Figure 3 The page shows merchants their current food safety risk issues and corresponding learning strategies, allowing merchants to clearly understand their current food safety issues and further correction plans. Figure 3 Click the View Risk button on the page to enter Figure 4 The page shown.

[0073] Please refer to Figure 4 , which is the detailed information of food safety risks of the merchant stores provided in the embodiment of this application. Figure 4 On the page, merchants can link to the specific food safety risk details that exist during the month, and merchants can further improve the problems based on this details.

[0074] Figure 5 for Figure 4 Data information on detailed information on food safety risks. Figure 5 In the report, merchants can understand the incidence rate of food safety risks and the number of food safety incidents that occurred in the month.

[0075] Figure 6 for Figure 5 The level of food safety risk issues occurring in the merchants and the corresponding handling methods for the food safety risk levels.

[0076] At the same time, the service end 102 will also provide the merchant end 103 used by the merchant with a learning course on how to handle the type of food safety issue, so that the merchant can handle the food safety issue in a timely manner and prevent further occurrence of food safety issues.

[0077] Figure 7 The food safety risk learning method provided by the embodiment of the present application is formulated for merchants. Specifically, Figure 7 Taking the methods of preventing and controlling pests and foreign objects as an example, the detailed information on preventing and controlling pests and foreign objects, the corresponding regulatory requirements and preventive measures are explained in detail.

[0078] In addition, the service problem attribution method provided in the embodiment of the present application can also be applied to the scenario of recommending food to users. For example, a user logs in to an online food delivery platform and searches for relevant food keywords in the search box of the online food delivery platform. The online food delivery platform server uses the corresponding recommendation algorithm based on the user's search keywords to obtain food information corresponding to the search keywords, sends the obtained food information to the user terminal, and the user terminal displays the obtained food information.

[0079] When the server determines the corresponding recommended food based on the search keywords provided by the user, the factors referenced by its recommendation algorithm include multiple factors, such as the user's dining habit information, the food feature information of the food searched by the user, the user's geographical location information, and the food safety risk level corresponding to the merchant providing the food. The food safety risk level of the merchant can be obtained using the service problem attribution model in the service problem attribution method provided in this application.

[0080] In addition, the service problem attribution method provided in the embodiment of the present application can also be applied to the merchant information recommended on the user's homepage. For example, after the user logs in to the online food delivery platform, the merchant information displayed on the user's homepage is obtained in advance by the server through a recommendation algorithm. Among them, the reference factors of the recommendation algorithm include: the distance information between the merchant's geographical location and the user's geographical location, the user's historical merchant types, food types, and the merchant's food safety risk level. Among them, the merchant's food safety risk level can be obtained using the service problem attribution model in the service problem attribution method provided in this application.

[0081] Among them, the food safety risk level of the merchant can be obtained through the following methods:

[0082] The user provides evaluation information for the merchant's takeaway products, and the service problem attribution model is used to analyze the evaluation information provided by the user to determine the food safety issues in the evaluation information and the types of causes that caused the food safety issues. In other words, if the evaluation information provided by the user for the food he or she consumes contains food safety issues, the specific safety problem classification results corresponding to the food safety issues are determined based on the service problem attribution model. The food safety risk level of the merchant corresponding to the food is determined based on the determined food safety problem classification results.

[0083] In an embodiment of the present application, a service problem attribution model is used to analyze the classification results of food safety issues contained in user evaluation information. The evaluation information for the meal and the classification results of food safety issues corresponding to the food safety issues in the evaluation information are sent to the merchant that provides the meal to the user. On the one hand, the merchant's food safety awareness is enhanced, and the merchant is reminded of the seriousness of the food safety problem. On the other hand, the merchant is reminded of the specific food safety problems sent by the user during the consumption of the meal. Thirdly, the merchant is provided with methods to deal with such food safety problems to avoid such food safety problems from happening again in the future. Among them, if the merchant has too many types of food safety problems, the merchant is provided with certain mandatory change measures.

[0084] Based on the above steps, the online food delivery system can determine the food safety level of the merchant, and when recommending the corresponding merchant list to the user, determine whether to recommend the merchant to the user based on the merchant's food safety level.

[0085] The service problem attribution model mentioned in the above application scenarios analyzes user evaluation information to obtain the service problem classification results corresponding to the user evaluation information. Among them, the service problem attribution model is specifically analyzed by the following method:

[0086] First, the service problem attribution model determines the service problem classification result for the text data to be analyzed based on the text data to be analyzed and the service knowledge graph associated with the text data to be analyzed.

[0087] Among them, the service knowledge graph refers to the prior knowledge graph of the text data to be analyzed, which is pre-stored in the service problem attribution model.

[0088] The service knowledge graph includes at least the following types:

[0089] 1) The service knowledge graph is the attributed service knowledge graph, that is, the knowledge graph of the service information in the text data to be analyzed. For example, the text data to be analyzed is: stinky tofu is very smelly. Its corresponding service knowledge graph is: stinky tofu is a kind of food, and the smell attribute of this food is smelly.

[0090] When the text data to be analyzed is analyzed using traditional analysis methods, the analysis result is that the stinky tofu has a deterioration problem.

[0091] However, in the embodiment of the present application, when the service problem attribution model analyzes the text data to be analyzed, "Stinky tofu is very smelly", it will simultaneously combine the service knowledge graph corresponding to the text data to be analyzed, "Stinky tofu is a kind of food, and the smell attribute of the food is smelly", to analyze the classification result of the service problem corresponding to the text data to be analyzed. In this example, the information of "Stinky tofu is very smelly" and "Stinky tofu is a kind of food, and the smell attribute of the food is smelly" can determine that the problem reported in the text data to be analyzed does not belong to any of the attribution types in the attribution standard of the service problem attribution model, indicating that the problem reported by the text data to be analyzed does not belong to a service problem.

[0092] The service problem attribution model vectorizes the input information, and the output information is also represented by a vector. For example, a vector of 1 indicates that the service information of the text data to be analyzed corresponds to at least one attribution type in the attribution standard. A vector of 0 indicates that the service information of the text data to be analyzed does not correspond to any attribution type in the attribution standard.

[0093] For another example, the text data to be analyzed is: This snail noodle smells so bad. The traditional analysis method will determine that the snail noodle has a spoilage problem. However, in the embodiment of the present application, the service problem attribution model combines the service knowledge graph when analyzing the text data to be analyzed. Here, the service knowledge graph corresponding to the text data to be analyzed is: snail noodle is a commodity, and the odor attribute of the commodity is smelly. Therefore, the service problem attribution model combines the text data to be analyzed and its corresponding service knowledge graph for analysis, and obtains that the smell of snail noodle is the product attribute of the snail noodle itself, indicating that the problem fed back in the text data to be analyzed does not belong to any of the attribution types in the attribution criteria of the service problem attribution model.

[0094] 2) The service knowledge graph is an ingredient graph. For example, the text data to be analyzed is "Why is there black pepper in black pepper beef tenderloin?", and its corresponding service knowledge graph is "The ingredient list of black pepper beef tenderloin includes: black pepper and beef tenderloin."

[0095] When the traditional analysis method is used to analyze the text data to be analyzed, the analysis result is that foreign matter exists in the black pepper beef tenderloin. However, in the embodiment of the present application, when the service problem attribution model analyzes the user evaluation information "Why is there black pepper in the black pepper beef tenderloin", it will be combined with its corresponding service knowledge graph "The ingredient list of black pepper beef tenderloin includes: black pepper and beef tenderloin". It is thus determined that the black pepper in the black pepper beef tenderloin is one of the main ingredients of this dish. Therefore, it is explained that the problem fed back by the text data to be analyzed does not belong to any of the attribution types in the attribution standard in the service problem attribution model, and there is no service problem.

[0096] 3) The service knowledge graph is a recipe. For example, when the text data to be analyzed is "What kind of fish is pickled fish with pickled cabbage", its corresponding service knowledge graph is "Pickled fish with pickled cabbage is a recipe, and its main ingredients are pickled cabbage and fish".

[0097] When analyzing the text data to be analyzed using traditional analysis methods, the analysis result is that the pickled fish problem is a food spoilage problem. However, in the embodiments of the present application, when the service problem attribution model analyzes the text data to be analyzed "What kind of fish is pickled fish with pickled cabbage", it will combine its corresponding service knowledge graph "Pickled fish with pickled cabbage is a recipe, and its main ingredients are pickled cabbage and fish". It is thus determined that pickled fish with pickled cabbage is a dish made from ordinary fish and pickled cabbage, rather than the name of a fish species. Therefore, it shows that the problem reflected in the text data to be analyzed does not belong to any of the attribution types in the attribution criteria of the service problem attribution model, and there is no service problem.

[0098] 4) The service knowledge graph is a common sense graph. For example, when the text data to be analyzed is "Let the mouse do it", its corresponding service knowledge graph is "Do it yourself".

[0099] When analyzing the text data to be analyzed using traditional analysis methods, the analysis result is a mouse. However, in the embodiments of the present application, when the service problem attribution model analyzes the text data to be analyzed "Let the mouse do it", it will combine its corresponding service knowledge graph "Do it yourself". It is thus determined that the text data to be analyzed is a prompt from the user to the merchant that, compared to the previous meal service provided by the merchant to the user, the current service is relatively poor, but it does not belong to any of the attribution types in the attribution criteria of the service problem attribution model, and hopes that the merchant will do it yourself and improve the current service quality.

[0100] The above are several common types of service knowledge graphs. When the service problem attribution model analyzes the language expression logic of the text data to be analyzed, by combining the service knowledge graph corresponding to the text data to be analyzed, the accuracy of the language expression logic of the text data to be analyzed can be improved. Further, after obtaining the language expression logic of the text data to be analyzed based on the text data to be analyzed and the corresponding service knowledge graph of the text data to be analyzed, the accuracy of determining the service classification result corresponding to the text data to be analyzed is also improved.

[0101] The process by which the service problem attribution model uses the text data to be analyzed and the service knowledge graph associated with the text data to be analyzed to determine the service problem attribution result corresponding to the text data to be analyzed is as follows:

[0102] Input the text data to be analyzed into the service problem attribution model to obtain the embedding vector of the text data to be analyzed, which includes the text embedding vector corresponding to each text unit in the text data to be analyzed, the paragraph embedding vector of each text unit in the text data to be analyzed, and the position embedding vector of each text unit in the text data to be analyzed. Each text unit can be each character or each word. Figure 1 In, E 这 Represents the text embedding vector corresponding to the word "this", E A represents the paragraph embedding vector of the word "this" in the text data to be analyzed, and E1 represents the position embedding vector of the word "this" in the text data to be analyzed.

[0103] Then, based on the embedding vector of the text data to be analyzed, the service knowledge graph embedding vector associated with the text data to be analyzed is obtained, and the service knowledge graph embedding vector includes the text embedding vector corresponding to each text unit in the service knowledge graph, the paragraph embedding vector of each text unit in the service knowledge graph, and the position embedding vector of each text unit in the service knowledge graph.

[0104] The text data to be analyzed is embedded in a vector and a service knowledge graph embedded in a vector associated with the text data to be analyzed to generate a first embedding vector for the text data to be analyzed. The service problem attribution model analyzes the first embedding vector for the text data to be analyzed, determines feature information of the text data to be analyzed for analyzing the service problem classification result, and determines the service problem classification result corresponding to the text data to be analyzed based on the feature information of the text data to be analyzed.

[0105] The service problem analysis model determines the service problem classification result corresponding to the text data to be analyzed based on the feature information of the text data to be analyzed. Specifically, the service problem classification result corresponding to the text data to be analyzed is determined based on the feature information of the text data to be analyzed, for example Figure 2 The classification results of foreign matter, spoilage, foodborne diseases, expired, and undercooked food are shown. Among them, the process of determining the target service problem classification result is also called fixing the shared layer. Then, after fixing the shared layer, select the target classification layer corresponding to the feature information of the text data to be analyzed, that is, determine the target service problem sub-classification result of the text data to be analyzed. For example, Figure 2 The pin shown in the figure. The determined target service problem sub-classification result is the target service problem classification result of the text data to be analyzed.

[0106] Second, the service problem attribution model determines the service problem classification result for the text data to be analyzed based on the text data to be analyzed and the pinyin information of the text data to be analyzed.

[0107] Specifically, the text data to be analyzed is input into the service problem attribution model to obtain an embedding vector of the text data to be analyzed, which includes a text embedding vector corresponding to each text unit in the text data to be analyzed, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector of each text unit in the text data to be analyzed, and a position embedding vector of each text unit in the text data to be analyzed.

[0108] The embedding vector of the text data to be analyzed includes the pinyin embedding vector corresponding to each text unit. Therefore, the service problem attribution model converts the text data to be analyzed into an embedding vector, which can improve the phenomenon of typos and thus improve the accuracy of the language expression logic of analyzing the text data to be analyzed.

[0109] Third, the service problem attribution model determines the service problem classification result for the text data to be analyzed based on the text data to be analyzed, the pinyin information of the text data to be analyzed, and the service knowledge graph associated with the text data to be analyzed.

[0110] Input the text data to be analyzed into the service problem attribution model to obtain the embedding vector of the text data to be analyzed, which includes the text embedding vector corresponding to each text unit in the text data to be analyzed, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector of each text unit in the text data to be analyzed, and the position embedding vector of each text unit in the text data to be analyzed. Figure 1 In, E 饭 is the word embedding vector corresponding to the word "fan", E A is the paragraph embedding vector of the word "rice" in the text data to be analyzed of "There are cockroaches in this meal", E3 is the position embedding vector of the word "rice" in the text data to be analyzed of "There are cockroaches in this meal", and E fan It is the pinyin embedding vector corresponding to the word "fan".

[0111] According to the embedding vector of the text data to be analyzed, the service knowledge graph embedding vector associated with the text data to be analyzed is determined, and the service knowledge graph embedding vector includes: the text embedding vector corresponding to each text unit in the service knowledge graph, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector of each text unit in the service knowledge graph, and the position embedding vector of each text unit in the service knowledge graph.

[0112] When the service problem attribution model analyzes the text data to be analyzed, on the basis of determining the text embedding vector, paragraph embedding vector and position embedding vector of the text data to be analyzed, the pinyin embedding vector is combined to reduce the frequency of typos. At the same time, the language expression logic of the text data to be analyzed is analyzed in combination with the service knowledge graph associated with the text data to be analyzed. Adding the pinyin embedding vector and the service knowledge graph can improve the accuracy of analyzing the language expression logic of the text data to be analyzed, so that the service problem attribution model can extract feature data for analyzing the service problem classification results from the text data to be analyzed according to the language expression logic of the text data to be analyzed. According to the feature data, firstly, in the service problem major classification result list of the service problem attribution model, the service problem major classification result corresponding to the feature data is determined, and then, based on the determined service problem major classification result, the service problem minor classification result corresponding to the feature information is selected from the service problem minor classification list of the service problem major classification result. The service problem classification result corresponding to the text data to be analyzed is determined according to the determined target service problem major classification result and the target service problem minor classification result in the target service problem major classification result.

[0113] In addition, the service problem attribution model is trained by the following method:

[0114] The first step is to obtain a pre-trained model for analyzing the feature information of text data.

[0115] The pre-trained model can be obtained by training the traditional Bert (Bidirectional Encoder Representations from Transformers, deep bidirectional pre-trained encoder) model. The Bert model is a pre-trained language model that can be used in tasks such as question-answering systems, sentiment analysis, spam filtering, named entity recognition, and document clustering.

[0116] The following is an introduction to the basic concepts of the Bert model:

[0117] The input object of the Bert model is the original vector of each word in the text to be recognized, including: word vector (TokenEmbedding), paragraph vector (Segment Embedding) and position vector (Position Embedding). The word vector, paragraph vector and position vector are directly added and processed and then input into the Bert model. Among them, the paragraph vector refers to the use of embedded information to separate the context of the text information to be recognized, and the position vector refers to the position information of the word in the text to be recognized.

[0118] The Bert model includes at least three types of inputs and corresponds to three types of outputs. They are as follows:

[0119] 1) Single text classification: For example, sentiment analysis of articles. The Bert model adds a [CLS] symbol vector to the input paragraph vector and uses the output vector corresponding to the symbol as the final semantic representation of the entire article. Compared with other texts, this [CLS] symbol vector can more fairly integrate the language expression logic of each word.

[0120] 2) Sentence-to-text classification: For example, question answering system. Its input vector not only adds [CLS] symbol vector, but also adds [SEP] symbol vector for sentence segmentation.

[0121] 3) Sequence plus annotation classification: In this type of task, the output vector corresponding to each word is the annotation of the word, which can be understood as classification.

[0122] In the embodiment of the present application, the Transformer structure in the Bert model acquires bidirectional information through a multi-head attention mechanism and a masking mechanism to enhance the language expression logic of the text data to be recognized. The attention mechanism allows the neural network to focus on a part of the input. The multi-head attention mechanism linearly combines multiple enhanced semantic vectors of each word to finally obtain an enhanced semantic vector with the same length as the original word vector. For example, in the two sentences, "Is this set meal / very / delicious" and "Is this set meal / very / delicious", the semantics expressed by the combination of "eat" and "good" are different.

[0123] Traditional language models input a text sequence from left to right, or combine left-to-right and right-to-left training. The bidirectional information acquisition of the Bert model has an advantage in understanding the context compared to the unidirectional language model.

[0124] The masking mechanism refers to randomly masking or replacing any character or word in a sentence when the text to be recognized is input into the model, so that the model can predict the content of the masked or replaced part of the sentence through understanding the context. Then, the model determines the attribution classification model corresponding to the text to be recognized based on the emotion reflected in the context of the text to be recognized after recognizing the masked or replaced content.

[0125] In this application, compared with the traditional Bert model, the pre-trained model adds a pinyin embedding vector to the original text embedding vector, paragraph embedding vector and position embedding vector when converting the text data to be analyzed into an embedding vector to improve the error rate of text recognition. Then, the service knowledge graph is embedded in the embedding vector of the text data to be analyzed to improve the accuracy of the language expression logic of the text data to be analyzed. Therefore, the pre-trained model is used to obtain the language expression logic of the text data to be analyzed, analyze and determine the feature information of the text data to be analyzed, and provide a basis for the subsequent determination of the service problem classification results of the text data to be analyzed.

[0126] In the second step, the pre-trained model is adjusted according to the text data samples for the service information and the service problem classification result samples for the text data samples to obtain a large classification text data feature analysis model, which is used to analyze the feature information of the text data corresponding to the service problem classification result.

[0127] The training purpose of the above-mentioned pre-trained model is to obtain the characteristic information of the text data to be analyzed. Here, in order to train the pre-trained model on the parameters of the service problem classification results, text data samples for service information and service problem classification result samples for the text data samples are used to train the parameters of the pre-trained model.

[0128] The trained large-category text data feature analysis model obtains the large-category service problem classification results corresponding to the text data to be analyzed based on the text data to be analyzed. Figure 2 As shown, it is a schematic diagram of the service problem classification results for text information in the service problem attribution model provided in the embodiment of the present application. The service problem classification results can be foreign matter, spoilage, foodborne diseases, expired, undercooked food, etc.

[0129] The third step is to adjust the large classification text data feature analysis model based on the text data samples of service information and the service problem small classification result samples of the text data samples to obtain a service problem attribution model for analyzing the service problem classification results of the text data to be analyzed.

[0130] The large classification text data feature analysis model is used to analyze the large classification results of service problems of text data. On this basis, the large classification text data feature analysis model is further adjusted in parameters so that it can analyze the small classification results of service problems of text data.

[0131] In the second step, the pre-trained model is adjusted using the text data samples for the service information and the service problem classification result samples for the text data samples to obtain the large classification text data feature analysis model. The "service problem classification result samples for the text data samples" mentioned here refer to the service problem classification result samples belonging to the text data samples, which are used as positive samples.

[0132] In addition, the pre-training model may be trained based on a text data sample for service information and a service problem classification result sample that does not belong to the text data sample as a first negative sample pair.

[0133] For example, if the text data sample is "There is a cockroach in the meal", the service problem classification result sample belonging to the text data sample is "foreign matter", which is a positive sample. The service problem classification result sample that does not belong to the text data sample can be "spoilage", or "foodborne disease", or "expired", or "undercooked food", etc., which are negative samples.

[0134] Positive samples are samples of service problem classification results that belong to the text data samples that can be queried by the large classification text data feature analysis model after training, and negative samples are comparison samples used as positive samples, so that the large classification text data feature analysis model can identify service problem classification results that do not belong to the text data samples. The large classification text data feature analysis model is trained with positive and negative samples, so that the classification accuracy of the large classification text data sample feature analysis model for the service problem classification results corresponding to the text data is improved.

[0135] In the third step, the text data sample for the service information and the service problem sub-classification result sample for the text data sample are used to adjust the large classification text data feature analysis model to obtain the service problem attribution model. The "service problem sub-classification result sample for the text data sample" mentioned here refers to the service problem sub-classification result belonging to the text data sample, which is used as a positive sample. Moreover, the service problem sub-classification result belonging to the text data sample here is at least one of the multiple candidate service problem sub-classification results in the service problem large classification result determined in the second step.

[0136] In addition, the large classification text data feature analysis model may be trained based on the text data sample for service information and the service problem small classification result sample that does not belong to the text data sample as a second negative sample pair.

[0137] For example, if the text data sample is "There is a cockroach in the dish", the service problem large classification result sample belonging to the text data sample is "foreign matter", and the service problem small classification result sample belonging to the text data sample is "cockroach". The service problem small classification result sample that does not belong to the text data sample can be "sharp foreign matter" or "rat", etc., which are negative samples.

[0138] The large-category text data feature analysis model is trained using positive samples and negative samples, so that the classification accuracy of the service problem attribution model obtained for the service problem classification results corresponding to the text data is improved.

[0139] An embodiment of the present application provides a service problem attribution method, comprising: obtaining text data to be analyzed for service information; inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed; wherein the service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.

[0140] In one embodiment of the present application, the service problem attribution model analyzes the text data to be analyzed and the service knowledge graph related to the text data to be analyzed to obtain the service problem classification result for the text data to be analyzed. The service problem attribution model provided in the above method takes the text data to be analyzed and the service knowledge graph related to the text data to be analyzed as input information. The service knowledge graph can obtain the attribute information of the service information in the text data to be analyzed, thereby improving the understanding of the language expression logic of the text data to be analyzed. Based on the service knowledge graph, it is helpful to improve the accuracy of the service problem attribution model in determining the language expression logic of the text data to be analyzed, and the accuracy of the service problem attribution model in matching the text data to be analyzed with the service problem classification result is also improved.

[0141] First embodiment

[0142] In the first embodiment of the present application, a service problem attribution method is provided. The specific process is as follows: Figure 8 As shown, it is a flowchart of a service problem attribution method provided in the first embodiment of the present application. Figure 8 The service problem attribution method shown includes: step S801 to step S802.

[0143] like Figure 8 As shown, in step S801, text data to be analyzed for service information is obtained.

[0144] This step is used to obtain the text data to be analyzed for the service information. After the server obtains the text data to be analyzed, it is the data basis for the service problem attribution model in the subsequent steps to determine the service problem classification result corresponding to the text data to be analyzed.

[0145] Among them, the service information can be food service information for food services provided by merchants to users; the analysis text data for the service information can be evaluation information of the food service information by users; and obtaining the text data to be analyzed for the service information includes: obtaining the text data to be analyzed for the food service information sent by the user end.

[0146] For example, the service information takes the meals provided by the merchant to the user as an example, and the text data to be analyzed for the service information is the evaluation information of the user on the meals.

[0147] like Figure 8 As shown, in step S802, the text data to be analyzed is input into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed.

[0148] This step is used to obtain the service problem classification result corresponding to the text data to be analyzed according to the service problem attribution model. The following describes a first method for the service problem attribution model to obtain the service problem classification result for the text data to be analyzed according to the text data to be analyzed and the service knowledge graph related to the text data to be analyzed.

[0149] The service problem attribution model is used to obtain the service problem classification result for the text data to be analyzed based on the text data to be analyzed and the service knowledge graph related to the text data to be analyzed. The service knowledge graph helps to understand the language expression logic of the text data to be analyzed and improve the accuracy of analyzing the language expression logic of the text data to be analyzed. Therefore, the text data to be analyzed is input into the service problem attribution model, and the service problem attribution model determines the service problem classification result corresponding to the text data to be analyzed based on the text data to be analyzed and the service knowledge graph related to the text data to be analyzed.

[0150] It should be noted that the step of inputting the text data to be analyzed into the service problem attribution model to obtain the service problem classification result for the text data to be analyzed includes:

[0151] Step 1-1: input the text data to be analyzed into a service problem attribution model to obtain a first embedding vector for the text data to be analyzed, wherein the first embedding vector includes an embedding vector of the text data to be analyzed and an embedding vector of a service knowledge graph related to the text data to be analyzed;

[0152] Step 1-2: According to the first embedding vector, obtain the service problem classification result corresponding to the text data to be analyzed.

[0153] Among them, in step 1-1, the inputting of the text data to be analyzed into the service problem attribution model to obtain a first embedding vector for the text data to be analyzed includes: inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector of the text data to be analyzed; according to the embedding vector of the text data to be analyzed, querying a service knowledge graph embedding vector associated with the embedding vector of the text data to be analyzed; according to the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph, obtaining a first embedding vector for the text data to be analyzed.

[0154] It should be noted that the service problem attribution model obtains the text data to be analyzed, obtains the service knowledge graph related to the text data to be analyzed, takes the text data to be analyzed and the service knowledge graph together as input information, and analyzes and determines the service problem classification result of the text data to be analyzed. Among them, the service problem attribution model converts the data form of the text data to be analyzed into the embedding vector of the text data to be analyzed, and converts the data form of the service knowledge graph into the embedding vector of the service knowledge graph. According to the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph related to the text data to be analyzed, a first embedding vector for the text data to be analyzed is generated.

[0155] Among them, obtaining the service knowledge graph related to the text to be analyzed can be to query the service knowledge graph related to the text data to be analyzed from the service knowledge graph library pre-stored in the service problem attribution model, or to obtain the service knowledge graph related to the text data to be analyzed from other sharing platforms.

[0156] Wherein, the step of inputting the text data to be analyzed into the service problem attribution model to obtain the embedding vector of the text data to be analyzed includes: inputting the text data to be analyzed into the service problem attribution model to obtain the text embedding vector corresponding to each text unit in the text data to be analyzed, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed; processing the text embedding vector corresponding to each text unit in the text data to be analyzed, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed to obtain the embedding vector of the text data to be analyzed.

[0157] Among them, querying the service knowledge graph embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed includes: obtaining the service information embedding vector in the text data to be analyzed according to the text data embedding vector to be analyzed; querying the service knowledge graph for the service information according to the service information embedding vector in the text data to be analyzed; obtaining the text embedding vector corresponding to each text unit in the service knowledge graph for the service information, the paragraph embedding vector corresponding to each text unit in the service knowledge graph, and the position embedding vector corresponding to each text unit in the service knowledge graph; processing the text embedding vector corresponding to each text unit in the service knowledge graph for the service information, the paragraph embedding vector corresponding to each text unit in the service knowledge graph, and the position embedding vector corresponding to each text unit in the service knowledge graph to obtain the service knowledge graph embedding vector for the service information.

[0158] In step 1-2, according to the first embedding vector, the service problem classification result corresponding to the text data to be analyzed is obtained, which can be obtained in at least three ways, which are discussed below.

[0159] (I) According to the first embedding vector, obtaining the service problem classification result corresponding to the text data to be analyzed can be obtained in the following first manner:

[0160] According to the first embedding vector, characteristic information of the text data to be analyzed is obtained; according to the characteristic information of the text data to be analyzed, a target service problem classification result matching the characteristic information of the text data to be analyzed is queried in the service problem classification list of the service problem attribution model as the service problem classification result corresponding to the text data to be analyzed.

[0161] Among them, according to the feature information of the text data to be analyzed, searching the service problem classification list of the service problem attribution model for a target service problem classification result that matches the feature information of the text data to be analyzed as the service problem classification result corresponding to the text data to be analyzed, including: obtaining feature information corresponding to multiple candidate service problem classification results in the service problem classification list of the service problem attribution model; comparing the feature information of the text data to be analyzed with the feature information corresponding to the multiple candidate service problem classification results, and determining the candidate service problem classification result containing the feature information of the text data to be analyzed as the target service problem classification result that matches the feature information of the text data to be analyzed as the service problem classification result corresponding to the text data to be analyzed.

[0162] For example, the text data to be analyzed is “mildew spots appear on the fruits in this fruit drink”, and the feature information of the text data to be analyzed is determined to be “fruits and mildew spots” according to the first embedding vector.

[0163] The first embedding vector includes the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph, where the embedding vector of the text data to be analyzed includes a text unit embedding vector, a position embedding vector, and a paragraph embedding vector. The first embedding vector is an embedding vector obtained by combining the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph. Therefore, to analyze the feature information of the text data to be analyzed according to the first embedding vector, it is necessary to combine the semantics of the text data to be analyzed and the service knowledge graph corresponding to the text data to be analyzed to determine the comprehensive semantics of the text data to be analyzed.

[0164] Here, the service knowledge graph corresponding to the text data to be analyzed "The fruit in this fruit drink has mold spots on it" is "Fruits belong to fresh fruits and vegetables, and mold spots on the surface of fresh fruits and vegetables belong to moldy phenomena." Therefore, according to the first embedding vector, it can be known that the feature information of the text data to be analyzed is "fruits and mold spots."

[0165] There are multiple candidate service problem classification results in the service problem classification list, such as Figure 2 The classification results of the two candidate service problems "visible mold and deterioration" and "rot" are shown. Among them, the characteristic information of "visible mold and deterioration" includes "mold features appearing on the surface of the object", and the characteristic information of "rot" includes "a large number of lesions appearing inside the object". Comparing the characteristic information of the text data to be analyzed with the characteristic information of the above two candidate service problem classification results, the service problem classification result corresponding to the text data to be analyzed "mold spots appear on the fruit in this fruit drink" is visible mold and deterioration.

[0166] (ii) Obtaining the service problem classification result corresponding to the text data to be analyzed according to the first embedding vector may be obtained in the following second manner:

[0167] According to the first embedding vector, characteristic information of the text data to be analyzed is obtained; according to the characteristic information of the text data to be analyzed, at least one service problem classification result matching the characteristic information of the text data to be analyzed is searched in the service problem classification list of the service problem attribution model; and the at least one service problem classification result is used as the service problem classification result corresponding to the text data to be analyzed.

[0168] (III) Obtaining the service problem classification result corresponding to the text data to be analyzed according to the first embedding vector may be obtained in the following third manner:

[0169] According to the first embedding vector, characteristic information of the text data to be analyzed is obtained; according to the characteristic information of the text data to be analyzed, a service problem major classification result associated with the characteristic information of the text data to be analyzed is queried in a service problem major classification list in the service problem attribution classification model as a target service problem major classification result of the text data to be analyzed; on the basis of determining the target service problem major classification result of the text data to be analyzed, a service problem minor classification result associated with the characteristic information of the text data to be analyzed is queried in a service problem minor classification list of the target service problem major classification result; according to the target service problem major classification result and the target service problem minor classification result, a service problem classification result corresponding to the text data to be analyzed is generated.

[0170] Among them, the service problem attribution model includes a major classification text data feature analysis model, and the major classification text data feature analysis model is used to analyze the feature information of the text data corresponding to the service problem major classification result; according to the feature information of the text data to be analyzed, the service problem major classification result associated with the feature information of the text data to be analyzed is queried in the service problem major classification list in the service problem attribution classification model as the target service problem major classification result of the text data to be analyzed, including: according to the major classification text data feature analysis model, obtaining the first feature information of the text data used for analyzing the service problem major classification result in the text data to be analyzed; obtaining the first candidate feature information of the text data corresponding to each candidate service problem major classification result in the service problem major classification list; comparing the first feature information in the text data to be analyzed with the first candidate feature information corresponding to each candidate service problem major classification result, to obtain the target service problem major classification result corresponding to the text data to be analyzed.

[0171] Among them, on the basis of determining the target service problem major classification result of the text data to be analyzed, searching the service problem minor classification list of the target service problem major classification result for the target service problem minor classification result associated with the feature information of the text data to be analyzed, including: obtaining second feature information of the text data used to analyze the service problem minor classification result in the text data to be analyzed; obtaining second candidate feature information of the text data corresponding to each candidate service problem minor classification result in the service problem minor classification list; comparing the second feature information in the text data to be analyzed with the second candidate feature information corresponding to each candidate service problem minor classification result to obtain the target service problem minor classification result corresponding to the text data to be analyzed.

[0172] The above is the first method for the service problem attribution model to obtain the service problem classification result for the text data to be analyzed based on the text data to be analyzed and the service knowledge graph related to the text data to be analyzed. According to the text data to be analyzed and the service knowledge graph related to the text data to be analyzed, the accuracy of the language expression logic of the text data to be analyzed is improved, and on this basis, the accuracy of the service problem classification result corresponding to the text data to be analyzed is improved.

[0173] The first method mentioned above is based on the service knowledge graph to improve the accuracy of the language expression logic of the text data to be analyzed. In addition, the service problem attribution model obtains the service problem classification result corresponding to the text data to be analyzed, which also includes the second method:

[0174] The service problem attribution model is specifically used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed, the pinyin information corresponding to the text data to be analyzed, and the service knowledge graph related to the text data to be analyzed.

[0175] In the second method, the service problem classification results corresponding to the text data to be analyzed are analyzed not only by combining the service knowledge graph of the text data to be analyzed, but also by combining the pinyin information of the text data to be analyzed, so as to improve the typos and other problems of the text data to be analyzed, and also improve the accuracy of the language expression logic of the text data to be analyzed.

[0176] The step of inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes:

[0177] Step 2-1: input the text data to be analyzed into the service problem attribution model to obtain a second embedding vector for the text data to be analyzed, wherein the second embedding vector includes a pinyin embedding vector of the text data to be analyzed and a pinyin embedding vector of a service knowledge graph related to the text data to be analyzed;

[0178] Step 2-2: According to the second embedding vector, obtain the service problem classification result corresponding to the text data to be analyzed.

[0179] Wherein, in step 2-1, the inputting of the text data to be analyzed into the service problem attribution model to obtain a second embedding vector for the text data to be analyzed includes: inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector of the text data to be analyzed, the embedding vector of the text data to be analyzed including a pinyin embedding vector corresponding to each text unit in the text data to be analyzed; according to the embedding vector of the text data to be analyzed, querying a service knowledge graph embedding vector associated with the embedding vector of the text data to be analyzed, the service knowledge graph embedding vector including a pinyin embedding vector corresponding to each text unit in the service knowledge graph; according to the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph, obtaining a second embedding vector for the text data to be analyzed.

[0180] It should be noted that the service problem attribution model converts the data form of the text data to be analyzed into the form of an embedding vector. In addition to obtaining the text unit embedding vector, position embedding vector, and paragraph embedding vector of the text data to be analyzed, it also obtains the pinyin embedding vector of the text unit. Figure 1 The text data to be analyzed in "There are cockroaches in this meal" contains the word embedding vector, paragraph embedding vector, position embedding vector and pinyin embedding vector of each word. The addition of the pinyin embedding vector of the text unit can improve the probability of typos when recognizing the text data to be analyzed. Then, the embedding vector of the text data to be analyzed is combined with the embedding vector of the service knowledge graph related to the text data to be analyzed to generate a second embedding vector for the text data to be analyzed. In this process, the combination of the pinyin embedding vector and the service knowledge graph not only improves the accuracy of text recognition in the text data to be analyzed, but also improves the accuracy of analyzing the language logic expression of the text data to be analyzed.

[0181] Among them, the second embedding vector obtained in step 201 is obtained after vector processing of the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph. The embedding vector of the text data to be analyzed in this step includes the pinyin embedding vector of the text data to be analyzed.

[0182] The step of inputting the text data to be analyzed into the service problem attribution model to obtain the embedding vector of the text data to be analyzed includes: inputting the text data to be analyzed into the service problem attribution model to obtain the text embedding vector corresponding to each text unit in the text data to be analyzed, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed; processing the text embedding vector corresponding to each text unit in the text data to be analyzed, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed to obtain the embedding vector of the text data to be analyzed.

[0183] The method of querying the service knowledge graph embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed includes: obtaining the service information embedding vector in the text data to be analyzed according to the text data embedding vector to be analyzed; querying the service knowledge graph for the service information according to the service information embedding vector in the text data to be analyzed; obtaining the text embedding vector corresponding to each text unit in the service knowledge graph for the service information, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the service knowledge graph, and the position embedding vector corresponding to each text unit in the service knowledge graph; processing the text embedding vector corresponding to each text unit in the service knowledge graph for the service information, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the service knowledge graph, and the position embedding vector corresponding to each text unit in the service knowledge graph to obtain the service knowledge graph embedding vector for the service information.

[0184] The above is the service problem classification result of the text data to be analyzed obtained by the service problem attribution model through the second method. Specifically, when the service problem attribution model converts the text data to be analyzed into an embedding vector, on the basis of the originally obtained text embedding vector, paragraph embedding vector, and position embedding vector, it combines the pinyin embedding vector corresponding to each text unit to reduce the problem of typos in the text data to be analyzed. At the same time, combined with the service knowledge graph corresponding to the text data to be analyzed, the accuracy of the language expression logic of the text data to be analyzed is improved. On the basis of determining the language expression logic of the text data to be analyzed, the service problem classification result corresponding to the text data to be analyzed is obtained. In the multiple candidate service problem sub-classification lists of the determined service problem classification results, the service problem sub-classification results corresponding to the text data to be analyzed are continued to be matched, and the service problem classification results corresponding to the text data to be analyzed are generated according to the service problem classification results and the service problem sub-classification results.

[0185] An embodiment of the present application provides a service problem attribution method, comprising: obtaining text data to be analyzed for service information; inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed; wherein the service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.

[0186] In the above method, the service problem attribution model analyzes the text data to be analyzed and the service knowledge graph related to the text data to be analyzed to obtain the service problem classification results for the text data to be analyzed. The service problem attribution model provided in the above method takes the text data to be analyzed and the service knowledge graph related to the text data to be analyzed as input information. The service knowledge graph can obtain the attribute information of the service information in the text data to be analyzed, thereby improving the understanding of the language expression logic of the text data to be analyzed. Based on the service knowledge graph, it is helpful to improve the accuracy of the service problem attribution model in determining the language expression logic of the text data to be analyzed, and the accuracy of the service problem attribution model in matching the service problem classification results with the text data to be analyzed is also improved.

[0187] Second embodiment

[0188] Corresponding to the embodiment corresponding to the application scenario of the service problem attribution method provided by the present application and the service problem attribution method provided by the first embodiment, the second embodiment of the present application also provides a service problem attribution device. Since the device embodiment is basically similar to the embodiment corresponding to the application scenario and the first embodiment, the description is relatively simple. For relevant parts, please refer to the embodiment corresponding to the application scenario and the partial description of the first embodiment. The device embodiment described below is only illustrative.

[0189] Please refer to Fig. 9 , which is a schematic diagram of a service problem attribution device provided in the second embodiment of the present application. A service problem attribution device provided in the second embodiment of the present application includes:

[0190] The first obtaining unit 901 is used to obtain text data to be analyzed for service information.

[0191] The second obtaining unit 902 is used to input the text data to be analyzed into the service problem attribution model to obtain the service problem classification result for the text data to be analyzed; wherein the service problem attribution model is used to obtain the service problem classification result for the text data to be analyzed based on the text data to be analyzed and the service knowledge graph related to the text data to be analyzed.

[0192] Third embodiment

[0193] Corresponding to the embodiment corresponding to the application scenario of the service problem attribution method provided by the present application and the service problem attribution method provided by the first embodiment, the third embodiment of the present application also provides another service problem attribution method.

[0194] Please refer to Fig.10 , which is a schematic diagram of another service problem attribution method provided in the third embodiment of the present application. Fig.10 The service problem attribution method shown includes: step S1001 to step S1002.

[0195] like Fig.10 As shown, in step S1001, text data to be analyzed for service information is obtained.

[0196] like Fig.10 As shown, in step S1002, the text data to be analyzed is input into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed.

[0197] The service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and the pinyin information corresponding to the text data to be analyzed.

[0198] Optionally, the step of inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model to obtain an embedding vector of the text data to be analyzed, the embedding vector of the text data to be analyzed including a text embedding vector corresponding to each text unit in the text data to be analyzed, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a position embedding vector corresponding to each text unit in the text data to be analyzed; and obtaining a service problem classification result corresponding to the text data to be analyzed according to the embedding vector of the text data to be analyzed.

[0199] Optionally, the step of inputting the text data to be analyzed into the service problem attribution model to obtain the embedding vector of the text data to be analyzed includes: inputting the text data to be analyzed into the service problem attribution model, performing vectorization processing on the text data to be analyzed, and obtaining a text embedding vector corresponding to each text unit in the text data to be analyzed, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a position embedding vector corresponding to each text unit in the text data to be analyzed; processing the text embedding vector corresponding to each text unit in the text data to be analyzed, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed to obtain the embedding vector of the text data to be analyzed.

[0200] Optionally, obtaining a service problem classification result corresponding to the text data to be analyzed based on the embedding vector of the text data to be analyzed includes: obtaining feature information of the text data to be analyzed based on the embedding vector of the text data to be analyzed; and searching for a target service problem classification result that matches the feature information of the text data to be analyzed in a service problem classification list of the service problem attribution model based on the feature information of the text data to be analyzed, as the service problem classification result corresponding to the text data to be analyzed.

[0201] Optionally, obtaining a service problem classification result corresponding to the text data to be analyzed based on the embedding vector of the text data to be analyzed includes: obtaining feature information of the text data to be analyzed based on the embedding vector of the text data to be analyzed; querying a service problem major classification result associated with the feature information of the text data to be analyzed in a service problem major classification list in the service problem attribution classification model based on the feature information of the text data to be analyzed, as a target service problem major classification result of the text data to be analyzed; querying a target service problem minor classification result associated with the feature information of the text data to be analyzed in a service problem minor classification list of the target service problem major classification result on the basis of determining the target service problem major classification result of the text data to be analyzed; generating a service problem classification result corresponding to the text data to be analyzed based on the target service problem major classification result and the target service problem minor classification result.

[0202] Optionally, the service problem attribution model is specifically used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed, the pinyin information corresponding to the text data to be analyzed, and the service knowledge graph related to the text data to be analyzed.

[0203] Optionally, the step of inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model to obtain a second embedding vector for the text data to be analyzed, the second embedding vector including an embedding vector of the text data to be analyzed and an embedding vector of a service knowledge graph related to the text data to be analyzed; and obtaining a service problem classification result corresponding to the text data to be analyzed according to the second embedding vector.

[0204] Optionally, the step of inputting the text data to be analyzed into the service problem attribution model to obtain a second embedding vector for the text data to be analyzed comprises: inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector for the text data to be analyzed, the embedding vector for the text data to be analyzed including a pinyin embedding vector corresponding to each text unit in the text data to be analyzed; according to the embedding vector for the text data to be analyzed, querying a service knowledge graph embedding vector associated with the embedding vector for the text data to be analyzed, the service knowledge graph embedding vector including a pinyin embedding vector corresponding to each text unit in the service knowledge graph; and obtaining a second embedding vector for the text data to be analyzed according to the embedding vector for the text data to be analyzed and the service knowledge graph embedding vector.

[0205] Optionally, obtaining the service problem classification result corresponding to the text data to be analyzed according to the second embedding vector includes: obtaining feature information of the text data to be analyzed according to the second embedding vector; querying the service problem major classification result associated with the feature information of the text data to be analyzed in the service problem major classification list in the service problem attribution classification model according to the feature information of the text data to be analyzed as a target service problem major classification result of the text data to be analyzed; querying the target service problem minor classification result associated with the feature information of the text data to be analyzed in the service problem minor classification list of the target service problem major classification result on the basis of determining the target service problem major classification result of the text data to be analyzed; generating the service problem classification result corresponding to the text data to be analyzed according to the target service problem major classification result and the target service problem minor classification result.

[0206] Fourth embodiment

[0207] Corresponding to the embodiment corresponding to the application scenario of the service problem attribution method provided by the present application and the service problem attribution method provided by the third embodiment, the fourth embodiment of the present application also provides another service problem attribution device. Since the device embodiment is basically similar to the embodiment corresponding to the application scenario and the third embodiment, the description is relatively simple. For relevant parts, please refer to the embodiment corresponding to the application scenario and the partial description of the third embodiment. The device embodiment described below is only illustrative.

[0208] Please refer to Fig.11 , which is a schematic diagram of another service problem attribution device provided in the fourth embodiment of the present application. Another service problem attribution device provided in the fourth embodiment of the present application includes:

[0209] The third obtaining unit 1101 is used to obtain text data to be analyzed for service information.

[0210] The fourth obtaining unit 1102 is used to input the text data to be analyzed into the service problem attribution model to obtain the service problem classification result for the text data to be analyzed. The service problem attribution model is used to obtain the service problem classification result for the text data to be analyzed based on the text data to be analyzed and the pinyin information corresponding to the text data to be analyzed.

[0211] Fifth embodiment

[0212] Corresponding to the embodiment corresponding to the application scenario of the service problem attribution method provided in the present application and the service problem attribution method provided in the first embodiment, the fifth embodiment of the present application also provides a training method for a service problem attribution model.

[0213] Please refer to Fig.12 , which is a schematic diagram of a training method for a service problem attribution model provided in the fifth embodiment of the present application. Fig.12 The service problem attribution method shown includes: steps S1201 to S1203.

[0214] like Fig.12 As shown, in step S1201, a pre-trained model for analyzing feature information of text data is obtained.

[0215] like Fig.12 As shown, in step S1202, the pre-trained model is adjusted according to the text data samples for the service information and the service problem classification result samples for the text data samples to obtain a large classification text data feature analysis model, and the large classification text data feature analysis model is used to analyze the feature information of the text data corresponding to the service problem classification result.

[0216] like Fig.12 As shown, in step S1203, the large classification text data feature analysis model is adjusted according to the text data samples for service information and the service problem small classification result samples for the text data samples to obtain a service problem attribution model for analyzing the service problem classification results for the text data to be analyzed.

[0217] Optionally, the pre-trained model analyzes feature information of text data in the following manner: obtaining a text data embedding vector based on the text data, the text data embedding vector including a text embedding vector corresponding to each text unit in the text data, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector of each text unit in the text data, and a position embedding vector of each text unit in the text data; analyzing the feature information of the text data based on the text data embedding vector.

[0218] Optionally, it also includes: obtaining a service knowledge graph embedding vector associated with the text data based on the text data embedding vector, the service knowledge graph embedding vector including a text embedding vector corresponding to each text unit in the service knowledge graph associated with the text data, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector of each text unit in the service knowledge graph, and a position embedding vector of each text unit in the service knowledge graph; analyzing feature information of the text data based on the text data embedding vector includes: analyzing feature information of the text data based on the text data embedding vector and the service knowledge graph embedding vector associated with the text data.

[0219] Optionally, the pre-trained model is adjusted according to the text data sample for service information and the service problem classification result sample for the text data sample to obtain the large classification text data feature analysis model, including: inputting the text data sample for service information into the pre-trained model to obtain a first service problem classification result for the text data sample for the service information output by the pre-trained model; and adjusting the service problem classification parameters of the pre-trained model according to the similarity between the first service problem classification result and the service problem classification result sample for the text data sample to obtain the large classification text data feature analysis model.

[0220] Optionally, the step of inputting the text data sample for the service information into the pre-trained model to obtain a first major classification result of service problems for the text data sample for the service information output by the pre-trained model comprises: inputting the text data sample for the service information into the pre-trained model to obtain feature information of the text data sample; and obtaining a first major classification result of service problems for the text data sample for the service information output by the pre-trained model based on the feature information of the text data sample.

[0221] Optionally, the large classification text data feature analysis model is adjusted according to the text data sample for service information and the service problem sub-classification result sample for the text data sample to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed, including: inputting the text data for the service information into the large classification text data feature analysis model to obtain a second service problem sub-classification result for the text data sample output by the large classification text data feature analysis model; and adjusting the large classification text data feature analysis model according to the degree of similarity between the second service problem sub-classification result and the service problem sub-classification result sample to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed.

[0222] Optionally, it also includes: taking text data samples for service information and samples of large classification results of service problems that do not belong to the text data as a first negative sample pair, adjusting the pre-trained model, and obtaining a large classification text data feature analysis model for analyzing the feature information of text data corresponding to the large classification results of service problems; taking text data samples for service information and samples of small classification results of service problems that do not belong to the text data as a second negative sample pair, adjusting the large classification text data feature analysis model, and obtaining a service problem attribution model for analyzing the service problem classification results of the text data to be analyzed.

[0223] Sixth embodiment

[0224] Corresponding to the embodiment corresponding to the application scenario of the service problem attribution method provided in this application and the training method of the service problem attribution model provided in the fifth embodiment, the sixth embodiment of this application also provides a training device for a service problem attribution model. Since the device embodiment is basically similar to the embodiment corresponding to the application scenario and the fifth embodiment, the description is relatively simple. For relevant parts, please refer to the embodiment corresponding to the application scenario and the partial description of the fifth embodiment. The device embodiment described below is only illustrative.

[0225] Please refer to Fig.13 , which is a schematic diagram of a training device for a service problem attribution model provided in the sixth embodiment of the present application. A training device for a service problem attribution model provided in the sixth embodiment of the present application includes:

[0226] The pre-trained model obtaining unit 1301 is used to obtain a pre-trained model for analyzing feature information of text data.

[0227] The large classification text data feature analysis model acquisition unit 1302 is used to adjust the pre-trained model according to the text data samples for the service information and the service problem large classification result samples for the text data samples, so as to obtain the large classification text data feature analysis model, wherein the large classification text data feature analysis model is used to analyze the feature information of the text data corresponding to the service problem large classification result.

[0228] The service problem attribution model acquisition unit 1303 is used to adjust the large classification text data feature analysis model according to the text data sample for the service information and the service problem small classification result sample for the text data sample, so as to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed.

[0229] Seventh embodiment

[0230] Corresponding to the above method embodiments provided by the present application, the seventh embodiment of the present application further provides an electronic device. Since the seventh embodiment is basically similar to the above method embodiments provided by the present application, the description is relatively simple, and the relevant parts can be referred to the description of the above method embodiments provided by the present application. The seventh embodiment described below is only illustrative.

[0231] Please refer to Fig.14, which is a schematic diagram of an electronic device provided in the seventh embodiment of the present application. The electronic device includes: at least one processor 1401, at least one communication interface 1402, at least one memory 1403 and at least one communication bus 1404; optionally, the communication interface 1402 can be an interface of a communication module, such as an interface of a WLAN (Wireless Local Area Network) module; the processor 1401 may be a processor CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application. The memory 1403 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Among them, the memory 1403 stores a program, and the processor 1401 calls the program stored in the memory 1403 to execute the above method provided in the embodiment of the present application.

[0232] Eighth embodiment

[0233] Corresponding to the above method embodiments provided in the present application, the eighth embodiment of the present application further provides a computer storage medium. Since the eighth embodiment is basically similar to the above method embodiments provided in the present application, the description is relatively simple. For relevant details, please refer to the description of the above method embodiments provided in the present application. The eighth embodiment described below is merely illustrative. The computer storage medium stores a computer program, and when the program is executed, the method provided in the above method embodiments is implemented. It should be noted that the detailed description of the storage medium provided in the eighth embodiment of the present application can refer to the relevant description of the above method embodiments provided in the present application, which will not be repeated here.

[0234] Although the present application is disclosed as above in the form of a preferred embodiment, it is not intended to limit the present application. Any technical personnel in this field may make possible changes and modifications without departing from the spirit and scope of the present application. Therefore, the scope of protection of the present application shall be based on the scope defined by the claims of the present application.

[0235] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory. The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0236] 1. Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be used to store information by any method or technology. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined in this article, computer-readable media does not include non-transitory media such as modulated data signals and carrier waves.

[0237] 2. Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems or computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

Claims

1. A service problem attribution method, characterized in that: include: Obtaining text data to be analyzed for service information, where the service information is meal information; Input the text data to be analyzed into the service problem attribution model to obtain the service problem classification result for the text data to be analyzed, including: obtaining the embedding vector of the text data to be analyzed; obtaining the embedding vector of the service information in the text data to be analyzed according to the embedding vector of the text data to be analyzed; querying the service knowledge graph for the service information according to the embedding vector of the service information in the text data to be analyzed; obtaining the embedding vector of the service knowledge graph for the service information; obtaining the first embedding vector for the text data to be analyzed according to the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph; obtaining the service problem classification result corresponding to the text data to be analyzed according to the first embedding vector; the service knowledge graph is a food ingredient graph or a recipe; the service problem classification result is a food safety problem type; The service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.

2. The method according to claim 1, characterized in that The first embedding vector includes an embedding vector of the text data to be analyzed and an embedding vector of a service knowledge graph related to the text data to be analyzed.

3. The method according to claim 1, characterized in that The step of obtaining the embedding vector of the text data to be analyzed includes: The text data to be analyzed is input into the service problem attribution model to obtain an embedding vector of the text data to be analyzed.

4. The method according to claim 1, characterized in that: The step of obtaining, according to the first embedding vector, a service problem classification result corresponding to the text data to be analyzed includes: Acquire feature information of the text data to be analyzed according to the first embedding vector; According to the feature information of the text data to be analyzed, a target service problem classification result matching the feature information of the text data to be analyzed is searched in the service problem classification list of the service problem attribution model as the service problem classification result corresponding to the text data to be analyzed.

5. The method according to claim 4, characterized in that The step of searching, according to the feature information of the text data to be analyzed, a target service problem classification result matching the feature information of the text data to be analyzed in the service problem classification list of the service problem attribution model as the service problem classification result corresponding to the text data to be analyzed comprises: Obtain feature information corresponding to a plurality of candidate service problem classification results in the service problem classification list of the service problem attribution model; The feature information of the text data to be analyzed is compared with the feature information corresponding to the multiple candidate service problem classification results, and the candidate service problem classification result containing the feature information of the text data to be analyzed is determined as the target service problem classification result that matches the feature information of the text data to be analyzed, as the service problem classification result corresponding to the text data to be analyzed.

6. The method according to claim 1, characterized in that The step of obtaining, according to the first embedding vector, a service problem classification result corresponding to the text data to be analyzed includes: Acquire feature information of the text data to be analyzed according to the first embedding vector; According to the feature information of the text data to be analyzed, searching the service problem classification list of the service problem attribution model for at least one service problem classification result that matches the feature information of the text data to be analyzed; The at least one service problem classification result is used as the service problem classification result corresponding to the text data to be analyzed.

7. The method according to claim 1, characterized in that The step of obtaining, according to the first embedding vector, a service problem classification result corresponding to the text data to be analyzed includes: Acquire feature information of the text data to be analyzed according to the first embedding vector; According to the feature information of the text data to be analyzed, in the service problem major classification list in the service problem attribution classification model, a service problem major classification result associated with the feature information of the text data to be analyzed is searched as a target service problem major classification result of the text data to be analyzed; On the basis of determining the target service problem major classification result of the text data to be analyzed, searching the service problem minor classification result associated with the feature information of the text data to be analyzed in the service problem minor classification list of the target service problem major classification result; According to the target service problem major classification result and the target service problem minor classification result, a service problem classification result corresponding to the text data to be analyzed is generated.

8. The method according to claim 7, characterized in that The service problem attribution model includes a large classification text data feature analysis model, and the large classification text data feature analysis model is used to analyze feature information of text data corresponding to the large classification results of the service problem; The step of searching, based on the feature information of the text data to be analyzed, a service problem classification result associated with the feature information of the text data to be analyzed in a service problem classification list in the service problem attribution classification model as a target service problem classification result of the text data to be analyzed includes: According to the large classification text data feature analysis model, first feature information of text data used for analyzing the large classification result of the service problem in the text data to be analyzed is obtained; Obtaining first candidate feature information of text data corresponding to each candidate service problem major classification result in the service problem major classification list; The first feature information in the text data to be analyzed is compared with the first candidate feature information corresponding to each candidate service problem classification result to obtain the target service problem classification result corresponding to the text data to be analyzed.

9. The method according to claim 7, characterized in that: The step of searching, based on determining the target service problem major classification result of the text data to be analyzed, the target service problem minor classification result associated with the feature information of the text data to be analyzed in the service problem minor classification list of the target service problem major classification result comprises: Acquire second feature information of text data used for analyzing the service problem sub-classification result in the text data to be analyzed; Acquire second candidate feature information of text data corresponding to each candidate service problem sub-classification result in the service problem sub-classification list; The second feature information in the text data to be analyzed is compared with the second candidate feature information corresponding to each candidate service problem sub-classification result to obtain the target service problem sub-classification result corresponding to the text data to be analyzed.

10. The method according to claim 3, characterized in that The step of inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector of the text data to be analyzed includes: Input the text data to be analyzed into the service problem attribution model to obtain a text embedding vector corresponding to each text unit in the text data to be analyzed, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a position embedding vector corresponding to each text unit in the text data to be analyzed; The text embedding vector corresponding to each text unit in the text data to be analyzed, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed are processed to obtain the embedding vector of the text data to be analyzed.

11. The method according to claim 1, characterized in that: The obtaining of a service knowledge graph embedding vector for the service information includes: Obtain a text embedding vector corresponding to each text unit in the service knowledge graph for the service information, a paragraph embedding vector corresponding to each text unit in the service knowledge graph, and a position embedding vector corresponding to each text unit in the service knowledge graph; The text embedding vector corresponding to each text unit in the service knowledge graph of the service information, the paragraph embedding vector corresponding to each text unit in the service knowledge graph, and the position embedding vector corresponding to each text unit in the service knowledge graph are processed to obtain the service knowledge graph embedding vector for the service information.

12. The method according to claim 1, characterized in that The service information is food service information for food services provided by merchants to users; The analysis text data for the service information is the user's evaluation information for the food service information; The obtaining of the text data to be analyzed for the service information includes: obtaining the text data to be analyzed for the food service information sent by the user end.

13. The method according to claim 1, characterized in that The service problem attribution model is specifically used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed, the pinyin information corresponding to the text data to be analyzed, and the service knowledge graph related to the text data to be analyzed.

14. The method according to claim 13, characterized in that The step of inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: Inputting the text data to be analyzed into a service problem attribution model to obtain a second embedding vector for the text data to be analyzed, wherein the second embedding vector includes a pinyin embedding vector of the text data to be analyzed and a pinyin embedding vector of a service knowledge graph related to the text data to be analyzed; According to the second embedding vector, a service problem classification result corresponding to the text data to be analyzed is obtained.

15. The method according to claim 14, characterized in that The step of inputting the text data to be analyzed into a service problem attribution model to obtain a second embedding vector for the text data to be analyzed includes: Inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector of the text data to be analyzed, wherein the embedding vector of the text data to be analyzed includes a pinyin embedding vector corresponding to each text unit in the text data to be analyzed; According to the embedding vector of the text data to be analyzed, querying the service knowledge graph embedding vector associated with the embedding vector of the text data to be analyzed, wherein the service knowledge graph embedding vector includes the pinyin embedding vector corresponding to each text unit in the service knowledge graph; According to the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph, a second embedding vector for the text data to be analyzed is obtained.

16. The method according to claim 15, characterized in that The step of inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector of the text data to be analyzed includes: Input the text data to be analyzed into the service problem attribution model, and obtain the text embedding vector corresponding to each text unit in the text data to be analyzed, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed; The text embedding vector corresponding to each text unit in the text data to be analyzed, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed are processed to obtain the embedding vector of the text data to be analyzed.

17. The method according to claim 15, characterized in that The querying, based on the embedding vector of the text data to be analyzed, a service knowledge graph embedding vector associated with the embedding vector of the text data to be analyzed includes: According to the embedding vector of the text data to be analyzed, obtaining the embedding vector of the service information in the text data to be analyzed; According to the service information embedding vector in the text data to be analyzed, querying the service knowledge graph for the service information; Obtain a text embedding vector corresponding to each text unit in the service knowledge graph for the service information, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector corresponding to each text unit in the service knowledge graph, and a position embedding vector corresponding to each text unit in the service knowledge graph; The text embedding vector corresponding to each text unit in the service knowledge graph of the service information, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the service knowledge graph, and the position embedding vector corresponding to each text unit in the service knowledge graph are processed to obtain the service knowledge graph embedding vector for the service information.

18. A service problem attribution method, characterized in that: include: Obtaining text data to be analyzed for service information, where the service information is meal information; Input the text data to be analyzed into the service problem attribution model to obtain the service problem classification result for the text data to be analyzed, including: obtaining the embedding vector of the text data to be analyzed; obtaining the embedding vector of the service information in the text data to be analyzed according to the embedding vector of the text data to be analyzed; querying the service knowledge graph for the service information according to the embedding vector of the service information in the text data to be analyzed; obtaining the embedding vector of the service knowledge graph for the service information; obtaining the first embedding vector for the text data to be analyzed according to the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph; obtaining the service problem classification result corresponding to the text data to be analyzed according to the first embedding vector; the service knowledge graph is a food ingredient graph or a recipe; the service problem classification result is a food safety problem type; The service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and the pinyin information corresponding to the text data to be analyzed.

19. The method according to claim 18, characterized in that The step of obtaining the embedding vector of the text data to be analyzed includes: The text data to be analyzed is input into the service problem attribution model to obtain an embedding vector of the text data to be analyzed, wherein the embedding vector of the text data to be analyzed includes a text embedding vector corresponding to each text unit in the text data to be analyzed, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a position embedding vector corresponding to each text unit in the text data to be analyzed.

20. The method according to claim 19, characterized in that The step of inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector of the text data to be analyzed includes: Input the text data to be analyzed into the service problem attribution model, perform vectorization processing on the text data to be analyzed, and obtain a text embedding vector corresponding to each text unit in the text data to be analyzed, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a position embedding vector corresponding to each text unit in the text data to be analyzed; The text embedding vector corresponding to each text unit in the text data to be analyzed, the pinyin embedding vector corresponding to each text unit, the paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and the position embedding vector corresponding to each text unit in the text data to be analyzed are processed to obtain the embedding vector of the text data to be analyzed.

21. The method according to claim 18, characterized in that The step of obtaining, according to the first embedding vector, a service problem classification result corresponding to the text data to be analyzed includes: Acquire feature information of the text data to be analyzed according to the first embedding vector; According to the feature information of the text data to be analyzed, a target service problem classification result matching the feature information of the text data to be analyzed is searched in the service problem classification list of the service problem attribution model as the service problem classification result corresponding to the text data to be analyzed.

22. The method according to claim 18, characterized in that The step of obtaining, according to the first embedding vector, a service problem classification result corresponding to the text data to be analyzed includes: Acquire feature information of the text data to be analyzed according to the first embedding vector; According to the feature information of the text data to be analyzed, in the service problem major classification list in the service problem attribution classification model, a service problem major classification result associated with the feature information of the text data to be analyzed is searched as a target service problem major classification result of the text data to be analyzed; On the basis of determining the target service problem major classification result of the text data to be analyzed, searching the service problem minor classification result associated with the feature information of the text data to be analyzed in the service problem minor classification list of the target service problem major classification result; According to the target service problem major classification result and the target service problem minor classification result, a service problem classification result corresponding to the text data to be analyzed is generated.

23. The method according to claim 18, characterized in that The service problem attribution model is specifically used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed, the pinyin information corresponding to the text data to be analyzed, and the service knowledge graph related to the text data to be analyzed.

24. The method according to claim 23, characterized in that The step of inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: Inputting the text data to be analyzed into a service problem attribution model to obtain a second embedding vector for the text data to be analyzed, wherein the second embedding vector includes an embedding vector of the text data to be analyzed and an embedding vector of a service knowledge graph related to the text data to be analyzed; According to the second embedding vector, a service problem classification result corresponding to the text data to be analyzed is obtained.

25. The method according to claim 24, characterized in that The step of inputting the text data to be analyzed into a service problem attribution model to obtain a second embedding vector for the text data to be analyzed includes: Inputting the text data to be analyzed into the service problem attribution model to obtain an embedding vector of the text data to be analyzed, wherein the embedding vector of the text data to be analyzed includes a pinyin embedding vector corresponding to each text unit in the text data to be analyzed; According to the embedding vector of the text data to be analyzed, querying the service knowledge graph embedding vector associated with the embedding vector of the text data to be analyzed, wherein the service knowledge graph embedding vector includes the pinyin embedding vector corresponding to each text unit in the service knowledge graph; According to the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph, a second embedding vector for the text data to be analyzed is obtained.

26. The method according to claim 24, characterized in that The step of obtaining, according to the second embedding vector, a service problem classification result corresponding to the text data to be analyzed includes: Acquire feature information of the text data to be analyzed according to the second embedding vector; According to the feature information of the text data to be analyzed, in the service problem major classification list in the service problem attribution classification model, a service problem major classification result associated with the feature information of the text data to be analyzed is searched as a target service problem major classification result of the text data to be analyzed; On the basis of determining the target service problem major classification result of the text data to be analyzed, searching the service problem minor classification result associated with the feature information of the text data to be analyzed in the service problem minor classification list of the target service problem major classification result; According to the target service problem major classification result and the target service problem minor classification result, a service problem classification result corresponding to the text data to be analyzed is generated.

27. A training method for a service problem attribution model, characterized in that: include: Obtain a pre-trained model for analyzing feature information of text data; According to the text data samples for the service information and the service problem classification result samples for the text data samples, the pre-trained model is adjusted to obtain a large classification text data feature analysis model, wherein the large classification text data feature analysis model is used to analyze feature information of the text data corresponding to the service problem classification result, wherein the service information is food information; According to the text data samples for service information and the service problem sub-classification result samples for the text data samples, the large classification text data feature analysis model is adjusted to obtain a service problem attribution model for analyzing the service problem classification results for the text data to be analyzed, wherein the service problem classification results are food safety problem types; The service problem attribution model is specifically used to: obtain an embedding vector of text data to be analyzed for service information; obtain an embedding vector of service information in the text data to be analyzed based on the embedding vector of text data to be analyzed; query a service knowledge graph for the service information based on the embedding vector of service information in the text data to be analyzed; obtain an embedding vector of the service knowledge graph for the service information; obtain a first embedding vector for the text data to be analyzed based on the embedding vector of text data to be analyzed and the embedding vector of the service knowledge graph; obtain a service problem classification result corresponding to the text data to be analyzed based on the first embedding vector; the service knowledge graph is a food ingredient graph or a recipe.

28. The method according to claim 27, characterized in that The pre-trained model analyzes the feature information of text data in the following way: Obtaining a text data embedding vector according to the text data, the text data embedding vector comprising a text embedding vector corresponding to each text unit in the text data, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector of each text unit in the text data, and a position embedding vector of each text unit in the text data; According to the text data embedding vector, feature information of the text data is analyzed.

29. The method according to claim 28, characterized in that Also includes: Obtaining a service knowledge graph embedding vector associated with the text data according to the text data embedding vector, wherein the service knowledge graph embedding vector includes a text embedding vector corresponding to each text unit in the service knowledge graph associated with the text data, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector of each text unit in the service knowledge graph, and a position embedding vector of each text unit in the service knowledge graph; The step of analyzing feature information of the text data according to the text data embedding vector comprises: The feature information of the text data is analyzed according to the text data embedding vector and the service knowledge graph embedding vector associated with the text data.

30. The method according to claim 27, characterized in that The method of adjusting the pre-trained model according to the text data samples of the service information and the service problem classification result samples of the text data samples to obtain the large classification text data feature analysis model includes: Inputting the text data sample for the service information into the pre-trained model to obtain a first service problem classification result for the text data sample for the service information output by the pre-trained model; According to the similarity between the first service problem classification result and the service problem classification result sample for the text data sample, the service problem classification parameters of the pre-trained model are adjusted to obtain the large classification text data feature analysis model.

31. The method according to claim 30, characterized in that The step of inputting the text data sample for the service information into the pre-trained model to obtain a first service problem classification result for the text data sample for the service information output by the pre-trained model includes: Inputting the text data sample for the service information into the pre-trained model to obtain feature information of the text data sample; A first service problem classification result for the text data sample of the service information output by the pre-training model is obtained according to the feature information of the text data sample.

32. The method according to claim 27, characterized in that The method of adjusting the large classification text data feature analysis model according to the text data sample for the service information and the service problem small classification result sample for the text data sample to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed includes: Inputting the text data for the service information into the large classification text data feature analysis model to obtain a second service problem sub-classification result for the text data sample output by the large classification text data feature analysis model; According to the similarity between the second service problem sub-classification result and the service problem sub-classification result sample, the large classification text data feature analysis model is adjusted to obtain a service problem attribution model for analyzing the service problem classification results for the text data to be analyzed.

33. The method according to claim 27, characterized in that Also includes: Taking the text data sample for the service information and the service problem classification result sample that does not belong to the text data as the first negative sample pair, adjusting the pre-trained model, and obtaining a large classification text data feature analysis model for analyzing feature information of the text data corresponding to the service problem classification result; The text data sample for service information and the service problem sub-classification result sample that does not belong to the text data are used as the second negative sample pair, and the large classification text data feature analysis model is adjusted to obtain a service problem attribution model for analyzing the service problem classification results of the text data to be analyzed.

34. A service problem attribution device, characterized in that: include: A first obtaining unit, configured to obtain text data to be analyzed for service information, wherein the service information is meal information; The second obtaining unit is used to input the text data to be analyzed into the service problem attribution model to obtain the service problem classification result for the text data to be analyzed, including: obtaining the embedding vector of the text data to be analyzed; obtaining the embedding vector of the service information in the text data to be analyzed according to the embedding vector of the text data to be analyzed; querying the service knowledge graph for the service information according to the embedding vector of the service information in the text data to be analyzed; obtaining the embedding vector of the service knowledge graph for the service information; obtaining the first embedding vector for the text data to be analyzed according to the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph; obtaining the service problem classification result corresponding to the text data to be analyzed according to the first embedding vector; the service knowledge graph is a food ingredient graph or a recipe; the service problem classification result is a food safety problem type; The service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.

35. A service problem attribution device, characterized in that: include: A third obtaining unit is used to obtain text data to be analyzed for service information, where the service information is food information; The fourth obtaining unit is used to input the text data to be analyzed into the service problem attribution model to obtain the service problem classification result for the text data to be analyzed, including: obtaining the embedding vector of the text data to be analyzed; obtaining the embedding vector of the service information in the text data to be analyzed according to the embedding vector of the text data to be analyzed; querying the service knowledge graph for the service information according to the embedding vector of the service information in the text data to be analyzed; obtaining the embedding vector of the service knowledge graph for the service information; obtaining the first embedding vector for the text data to be analyzed according to the embedding vector of the text data to be analyzed and the embedding vector of the service knowledge graph; obtaining the service problem classification result corresponding to the text data to be analyzed according to the first embedding vector; the service knowledge graph is a food ingredient graph or a recipe; the service problem classification result is a food safety problem type; The service problem attribution model is used to obtain a service problem classification result for the text data to be analyzed based on the text data to be analyzed and the pinyin information corresponding to the text data to be analyzed.

36. A training device for a service problem attribution model, characterized in that: include: A pre-trained model obtaining unit, used to obtain a pre-trained model for analyzing feature information of text data; A large classification text data feature analysis model acquisition unit is used to adjust the pre-trained model according to the text data samples for service information and the service problem large classification result samples for the text data samples to obtain a large classification text data feature analysis model, wherein the large classification text data feature analysis model is used to analyze feature information of text data corresponding to the service problem large classification result, wherein the service information is food information; A service problem attribution model obtaining unit is used to adjust the large classification text data feature analysis model according to the text data sample for the service information and the service problem small classification result sample for the text data sample, so as to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed, wherein the service problem classification result is a food safety problem type; The service problem attribution model is specifically used to: obtain an embedding vector of text data to be analyzed for service information; obtain an embedding vector of service information in the text data to be analyzed based on the embedding vector of text data to be analyzed; query a service knowledge graph for the service information based on the embedding vector of service information in the text data to be analyzed; obtain an embedding vector of the service knowledge graph for the service information; obtain a first embedding vector for the text data to be analyzed based on the embedding vector of text data to be analyzed and the embedding vector of the service knowledge graph; obtain a service problem classification result corresponding to the text data to be analyzed based on the first embedding vector; the service knowledge graph is a food ingredient graph or a recipe.

37. An electronic device, characterized in that: The electronic device comprises a processor and a memory; The memory stores a computer program, and after the processor runs the computer program, it executes the method described in any one of claims 1 to 33.

38. A computer storage medium, characterized in that The computer storage medium stores a computer program, and when the computer program is executed, the method according to any one of claims 1 to 33 is performed.

Images