Problem prediction method and electronic equipment

Abstract

The embodiment of the invention discloses a problem prediction method and electronic equipment, and the method comprises the steps: responding to a service request initiated by a user, and determining an associated target commodity transaction order; obtaining target data related to problem classification prediction of the target commodity transaction order; determining input data of an artificial intelligence (AI) generation model according to the obtained target data, so that the AI generation model generates an appeal text of a user or a to-be-consulted question text based on the target data; and inputting the user appeal text or the question text generated by the AI generation model into a preset classification prediction model so as to convert the user appeal text or the question text into a standardized question description text in a candidate set, and providing the standardized question description text to a client for a user to select. According to the embodiment of the invention, the problem prediction accuracy can be improved.

Description

Technical Field

[0001] The present application relates to the field of information processing technology, and in particular to a problem prediction method and an electronic device. Background Art

[0002] In the commodity information service system, the customer service system is a very important component. After placing an order, if the user encounters problems in logistics, commodity quality, etc., they can provide feedback through the customer service system to seek solutions. Among them, after the user initiates a customer service request based on an order, in order to be able to provide services to users more through the "customer service robot", the user is usually required to describe his problem as accurately or in a standardized manner as possible. In order to help users describe their problems, the questions that users need to ask can also be predicted and displayed to users. For example, "Guess what you want to ask" can be provided in the customer service dialogue interface, and the predicted questions that users may ask can be displayed in the interface. If the user does want to ask a certain other question, he can click on the question directly without having to type in the questions he wants to ask, thereby helping users improve efficiency and experience.

[0003] In the above-mentioned "Guess What You Want to Ask" service, the accuracy of problem prediction is relatively critical. In the prior art, one prediction method is to obtain the user's problem order, query some status of the order, such as refund information and logistics information, and then input this information into the model used for problem prediction to predict the order problems that the user may encounter. However, this method often predicts problems with low accuracy, which means that users still need to manually type their own questions in most cases, resulting in a poor user experience. In addition, the questions manually entered by users may inevitably be substandard, resulting in a large number of transfers to manual customer service, which in turn makes the system's customer service costs relatively high. Summary of the invention

[0004] The present application provides a problem prediction method and an electronic device, which can improve the accuracy of problem prediction.

[0005] This application provides the following solutions:

[0006] A problem prediction method, comprising:

[0007] In response to a service request initiated by a user, determining an associated target commodity transaction order;

[0008] Acquire target data related to problem classification prediction of the target commodity transaction order;

[0009] Determine input data of an artificial intelligence (AI) generation model according to the acquired target data, so that the AI ​​generation model generates a user's demand text or a question text to be consulted based on the target data;

[0010] The user demand text or question text generated by the AI ​​generation model is input into a preset classification prediction model to be converted into a standardized question description text in a candidate set, and the standardized question description text is provided to the client for user selection.

[0011] The step of determining the associated target commodity transaction order includes:

[0012] If the user initiates a customer service request after selecting an order, the order selected by the user is determined as the target commodity transaction order.

[0013] The step of determining the associated target commodity transaction order includes:

[0014] If the user has not selected an order before initiating a customer service request, a prediction of the problematic order is performed based on the status of each order in the order list associated with the user, and the target commodity transaction order is determined based on the prediction result.

[0015] Among them, the target data includes: historical customer service record information related to the target commodity transaction order, browsing and clicking record information executed by the user in the current commodity information service system client before initiating the customer service request, and / or status information of the target commodity transaction order in multiple dimensions.

[0016] The historical customer service record information related to the target commodity transaction order includes: record information generated during the communication between the user and the merchant user, the manual customer service and / or the intelligent customer service on the system side regarding the target commodity transaction order.

[0017] The step of determining the input data of the artificial intelligence AI generation model according to the acquired target data includes:

[0018] The acquired target data is arranged and processed according to the time dimension to generate timeline trajectory information, and the timeline trajectory information is determined as input data for the AI ​​generation model.

[0019] Among them, it also includes:

[0020] In the process of arranging the acquired target data, key data related to problem prediction is screened out, and the timeline trajectory information is generated according to the screened out key data.

[0021] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of any of the methods described above.

[0022] An electronic device, comprising:

[0023] one or more processors; and

[0024] A memory associated with the one or more processors, the memory being used to store program instructions, wherein the program instructions, when read and executed by the one or more processors, execute the steps of any of the methods described above.

[0025] A computer program product comprises a computer program / computer executable instructions, wherein the computer program / computer executable instructions implement the steps of any of the aforementioned methods when executed by a processor in an electronic device.

[0026] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0027] Through the embodiments of the present application, the ability of the AI ​​generation model can be used to enable more and richer information to participate in the process of classifying and predicting user questions, and the AI ​​generation model first generates summary information of the appeal text or the text of the question to be consulted based on the acquired target data. After that, the generation result output by the AI ​​generation model is further input into the small model for problem prediction, so that the prediction information of the small model is more complete and comprehensive, and the problem description text with standardized expression is output, so that the final prediction accuracy can be further improved.

[0028] Among them, there can be multiple specific target data to be obtained. For example, by obtaining the user's historical consulting service records, the model can know the demands expressed by the user in the past, and can confirm whether the user's demands have been resolved and whether further follow-up is needed; by obtaining the user's click and browse records before coming online, the model can know what the user is currently concerned about. For example, if the user frequently clicks on logistics information before coming online, it can be inferred that the user has a high probability of consulting logistics-related issues; in addition, the status information of specific orders in multiple dimensions can also be obtained.

[0029] Of course, any product implementing the present application does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1 It is a schematic diagram of the system architecture provided by the embodiment of the present application;

[0032] Figure 2 is a flow chart of the method provided in the embodiment of the present application;

[0033] Figure 3 is a schematic diagram of a device provided in an embodiment of the present application;

[0034] Figure 4 It is a schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application belong to the scope of protection of this application.

[0036] First of all, it should be noted that the customer service in the embodiment of the present application can mainly be the customer service provided by the platform of the commodity information system for consumer users. When consumer users have questions about specific orders, if they communicate with the merchant to no avail, or the merchant fails to provide a satisfactory solution, they can seek help through the customer service provided by the platform. Of course, consumer users can also communicate directly with the customer service of the platform, etc. In this process, due to the large number of users corresponding to the platform, if all services are provided by manual customer service, a large number of manual customer service is required, and the cost will be very high. In addition, the service quality will also be limited by the ability of the manual customer service itself, and further, it may also cause users to wait for a long time. Therefore, for the customer service system of the platform, the above problems will also be dealt with by deploying an intelligent customer service system. Among them, in the process of communicating with users through intelligent customer service, users may have many forms when asking questions, and users may not be able to express the questions they need to consult clearly, which will cause difficulties for intelligent customer service to understand users' questions. Therefore, when serving users through intelligent customer service, it will rely more on the standardized expression of user demands or problems. In this scenario, it is particularly important to predict the questions that users may need to ask and guide users to express them in a standardized manner.

[0037] In addition, the inventors of the present application found in the process of implementing the present application that in the prior art, when using some models for problem classification prediction to predict problems, the reason for inaccurate predictions is mainly due to the following reasons:

[0038] 1. Failure to combine the user's historical service records leads to repeated useless guesses. In other words, before the user initiates a consultation on an order, he may have communicated with the merchant before, or with the platform's intelligent customer service, manual customer service, etc. If the historical communication is not taken into account, and the problem is predicted directly based on the current status of the order, a problem may be predicted repeatedly. For example, a user needs to modify the delivery address for an order. The merchant has communicated with the merchant before, and a satisfactory solution has been given to the user. Afterwards, the user may need to communicate again about the logistics delivery time of the order. However, when the model predicts the problem, it may repeatedly predict the problem of modifying the delivery address, etc.

[0039] 2. There is too little status information related to orders and users, and it is impossible to truly depict the current status of users and orders. The existing problem classification prediction model mainly makes predictions based on some simple status information of the order, such as logistics status, refund status, etc. However, in fact, there are more dimensional information in the entire life cycle of the order, including buyer information, seller information, whether a partial refund or return has been applied, etc. This information is helpful for problem classification prediction, but it has not been applied to problem classification prediction.

[0040] 3. Simply using the problem classification prediction model for prediction. Since this model is a "small model" and lacks model capabilities, this is also one of the reasons why the problem classification prediction results are not accurate enough.

[0041] In summary, in the solutions of the prior art, on the one hand, there is a large amount of information that is not fully utilized, and on the other hand, the model itself is not capable enough, which leads to inaccurate results of question classification prediction. Among them, the model capabilities include restrictions on the amount of input information and restrictions on data analysis capabilities. In other words, if more information is simply input into the above-mentioned question classification prediction model, the model is not capable of processing so much information. In addition, this model lacks the ability to understand natural language, and is more likely to make predictions through keyword matching, mapping according to preset rules, etc. Therefore, it cannot achieve the purpose of improving prediction accuracy.

[0042] In view of the above situation, in the embodiment of the present application, an implementation method is provided by combining AI (Artificial Intelligence) large-scale parameters with the aforementioned problem classification prediction model. Among them, the AI ​​large-scale parameter model can be referred to as a "big model"), which refers to a deep learning model containing massive parameters. This AI big model is large in scale, can store and process a large amount of information, and has the ability to understand and generate multimodal information, thereby achieving higher performance on multiple tasks. In the embodiment of the present application, historical service information, information on more dimensions of the current order, user browsing and clicking records, etc. can be used as input information of the above-mentioned AI big model, and the big model understands and summarizes the above information to summarize what the user's possible demands or problems are. Of course, since the content generated by the AI ​​big model has a certain randomness, for example, with the same input information and the same goal, the content generated by the AI ​​big model each time may be different in terms of expression, etc.; and the intelligent customer service system usually needs to provide solutions to standardized problems, that is, the output result of the problem prediction system needs to be a pre-defined problem in a certain problem candidate set. Therefore, it is also possible to combine the small model for problem classification prediction to convert the user's demands or problems summarized by the big model into more standardized problems. What is finally presented to the user may be the standardized expression of the problem after being converted by the small model, for the user to choose.

[0043] From the perspective of system architecture, see Figure 1 , the embodiments of the present application can improve the existing intelligent customer service system, which may include a client and a server. After the user initiates a customer service request through the client, the server can determine the target commodity transaction order and obtain a variety of target data related to problem prediction, for example, it may include historical customer service records, browsing click records, more dimensional order information, and so on. Then, by combining the large AI model with the small model, the effective use of richer information is achieved, and the accuracy of the problem prediction results is improved. After predicting the standardized problem description text, it can be provided to the client for display, and the user can complete the problem description by selecting the prediction result, thereby saving the time required to enter the problem text. In addition, it can also solve the problem that users find it difficult to clearly express their problems.

[0044] The specific implementation scheme provided in the embodiments of the present application is described in detail below.

[0045] First, the present application embodiment provides a problem prediction method, see Figure 2 , the method may include:

[0046] S201: In response to a service request initiated by a user, determining an associated target commodity transaction order.

[0047] In specific implementation, users can be provided with multiple ways to initiate customer service requests. For example, an "official customer service" entrance can be provided in pages such as the order details page, or users can also provide an "exclusive customer service" entrance on their personal homepage (for example, "My"), etc. Users can initiate customer service requests to the platform's customer service system through the above entrances.

[0048] If the user initiates a customer service request through a page such as the details page of an order, the customer service request is associated with the order information by default, so the order associated with the request can be directly determined as the target commodity transaction order. If the user initiates a customer service request through the entrance provided in the personal homepage, since the user has not yet selected an order before the user comes in, the user can first perform an order selection operation after the user comes in, or the order that may need to be consulted can be predicted based on the status of the order associated with the user, and at this time, the order can be determined as the target commodity transaction order, etc.

[0049] S202: Acquire target data related to the problem classification prediction of the target commodity transaction order.

[0050] After determining the target commodity transaction order associated with the current customer service request, the target data for problem prediction can be obtained. In an embodiment of the present application, the target data specifically obtained may include multiple types, for example, one of which may be historical customer service record information related to the target commodity transaction order, that is, before the user currently inquires about a certain order, he may have communicated with the order before, or, before consulting the customer service of the platform, he may have communicated with the merchant, and so on. The records generated during these communications can be used as the basis for current problem prediction, which is conducive to avoiding repeated predictions of the same problem or problems that have been communicated and have obtained satisfactory solutions, etc. Among them, this historical customer service record can exist in the form of conversation content between the user and the customer service, and can specifically be multiple conversation texts.

[0051] Another target data may be the browsing and clicking record information performed by the user in the current commodity information service system client before initiating the customer service request. That is to say, after the user opens the client of the specific commodity information service system, before initiating a customer service request to the platform customer service, the user may have performed some browsing and clicking operations, for example, may have clicked to view the logistics details of a certain order, or may have viewed the refund progress of a certain order, etc. These operation information is also helpful for problem prediction and can also be collected as input information for the model.

[0052] The third type of target data may be status information of a specific order in multiple dimensions. For example, in addition to logistics information, refund and return information, it may also include dispute status, partial refund status, and so on.

[0053] Of course, in the specific implementation, other data can also be obtained as target data, for example, it can also include the information of specific users, or it can also include the product information associated with the order, the amount information, the seller user information, etc. In short, all the information that can be viewed or queried during the service provided by the manual customer service can be obtained as the data basis for model prediction.

[0054] S203: Determine input data of an artificial intelligence (AI) generation model according to the acquired target data, so that the AI ​​generation model generates a user's demand text or a question text to be consulted based on the target data.

[0055] After obtaining the above target data, the input data of the AI ​​generation model can be determined according to the target data. Specifically, the obtained target data can be directly used as the input data of the AI ​​generation model, or, since the amount of target data may be relatively large, and the AI ​​generation model may also have certain restrictions on the length of the input data, in practical applications, the target data can also be screened to filter out key data related to the problem prediction of the current order, and then use this key data as the input data of the AI ​​generation model. For example, for historical service record information, it may include some conversation content that is not related to order problem consultation, including greeting content, chat content, etc., which can be filtered out. Or, for the user's click browsing history, some browsing click records related to key pages can be retained. For example, key pages can include order list pages, logistics details pages, etc., and the browsing click records of these pages can be retained. As for the browsing of goods that have not yet been purchased, since it is meaningless to the after-sales scenario, such click browsing records can be filtered out, etc. In this way, the interference of invalid data can be reduced, and it is also easier to meet the requirements of the AI ​​generation model for the length of input data.

[0056] In addition, before the relevant data is specifically input into the AI ​​generation model, the acquired relevant data can also be arranged and processed according to the time dimension to generate timeline trajectory information, and the timeline trajectory information is determined as the input data of the AI ​​generation model. That is to say, whether it is a historical service record, a click-and-browse record, or the status information of an order, there is corresponding time information. Therefore, such relevant data can be arranged into timeline trajectory information according to the time information. Among them, each type of relevant data can correspond to its own timeline trajectory information, for example, including the timeline trajectory of historical service records (that is, what conversation content was generated at what time point, etc.), the timeline trajectory of click-and-browse records (what pages were clicked and browsed at what time point), the timeline trajectory information of the order status (what states were generated at what time point), and so on. The above timeline trajectory information can also be expressed in the form of text, and then it can be constructed into a prompt text (Prompt) for dialogue with the AI ​​generation model, and input into the AI ​​generation model to guide the AI ​​generation model to perform the specified task. In the embodiment of the present application, the designated task is to understand the input content and summarize the demands / problems that the user may have.

[0057] Among them, the AI ​​generation model can be a large model of the "Wenshengwen" type, that is, it has the ability to understand and summarize the natural language based on the input text and generate new text content. In the embodiment of the present application, fine-tuning training and other methods can be performed on the basis of the general Wenshengwen large model, so that the specific AI generation model has the ability to understand the natural language of the input timeline trajectory information in the customer service scenario, and summarize the user's demands / problems.

[0058] In short, through the above AI generation model, it is possible to summarize the input historical customer service record information, click browsing history, order status information and other related data to summarize the user's possible demands or problems, which can also be expressed in the form of text. For example, "I want to apply for a refund for this order", etc.

[0059] S204: Input the user demand text or question text generated by the AI ​​generation model into a preset classification prediction model to convert it into a standardized question description text in a candidate set, and provide the standardized question description text to the client for user selection.

[0060] After the AI ​​generation model completes the summary of user demands / problems, the content generated by the AI ​​generation model may be random in terms of expression, that is, under the same input data and prompt method, the content generated by the same AI generation model each time may be the same or similar in meaning, but there may be differences in the wording, etc. For example, in the above example, the content generated by the AI ​​generation model may be "I want to apply for a refund for this order", or "I want to apply for a refund", etc. In the embodiment of the present application, it is usually necessary to output a standardized description of the problem. Therefore, after the AI ​​generation model completes the summary of user demands / problems, the content output by the AI ​​generation model can also be input into the classification prediction model, and the model converts the content output by the AI ​​generation model into a standardized problem in the candidate set, and provides the standardized problem to the client for user selection. Among them, the candidate set can include multiple problem description texts, corresponding to the standardized problems defined in the system. In this way, the final generated can be a standardized problem description text. For example, assuming that the user demand or question text to be consulted summarized by the AI ​​generation model is "I want to apply for a refund for this order", after being converted by the small model, the output standardized problem description text can be: "I want to apply for a refund", etc. This standardized question description text can then be displayed to the user as a prediction result, and the user can choose to use the prediction result to ask a question. Of course, if the prediction result is inaccurate, the user can still describe the actual question to be asked through keyboard input, voice input, etc.

[0061] If the user selects the recommended question description text to ask a question, the intelligent customer service can give priority to answering the user's question, including dialogue with the user, further clarifying the user's needs, and providing corresponding solutions, etc. The user can also apply to transfer to manual customer service, or the intelligent customer service will automatically transfer the user to manual customer service.

[0062] In short, through the embodiments of the present application, with the help of the ability of the AI ​​generation model, more and richer information can be involved in the process of classifying and predicting user questions, and the AI ​​generation model first summarizes the acquired target data to generate a user demand text or a text of the question to be consulted. After that, the generated results output by the AI ​​generation model are further input into the small model for problem prediction, so that the prediction information of the small model is more complete and comprehensive, and the problem description text with standardized expression is output, so that the final prediction accuracy can be further improved.

[0063] Among them, there can be multiple specific target data to be obtained. For example, by obtaining the user's historical consulting service records, the model can know the demands expressed by the user in the past, and can confirm whether the user's demands have been resolved and whether further follow-up is needed; by obtaining the user's click and browse records before coming online, the model can know what the user is currently concerned about. For example, if the user frequently clicks on logistics information before coming online, it can be inferred that the user has a high probability of consulting logistics-related issues; in addition, the status information of specific orders in multiple dimensions can also be obtained.

[0064] It should be noted that the embodiments of the present application may involve the use of user data. In actual applications, user-specific personal data can be used in the scheme described herein within the scope permitted by applicable laws and regulations, subject to the requirements of applicable laws and regulations of the country where the user is located (for example, with the user's explicit consent, effective notification to the user, etc.).

[0065] Corresponding to the above method embodiment, the present application embodiment also provides a problem prediction device, see Figure 3 , the device may include:

[0066] The transaction order determination unit 301 is used to determine the associated target commodity transaction order in response to the service request initiated by the user;

[0067] A target data acquisition unit 302 is used to acquire target data related to the problem classification prediction of the target commodity transaction order;

[0068] The large model generation unit 303 is used to determine the input data of the artificial intelligence AI generation model according to the acquired target data, so that the AI ​​generation model generates the user's demand text or the question text to be consulted based on the target data;

[0069] The small model conversion unit 304 is used to input the user demand text or question text generated by the AI ​​generation model into a preset classification prediction model to convert it into a standardized question description text in a candidate set, and provide the standardized question description text to the client for user selection.

[0070] Among them, transaction order confirmation can be used specifically for:

[0071] If the user initiates a customer service request after selecting an order, the order selected by the user is determined as the target commodity transaction order.

[0072] Alternatively, the transaction order determination may be used specifically for:

[0073] If the user has not selected an order before initiating a customer service request, a prediction of the problematic order is performed based on the status of each order in the order list associated with the user, and the target commodity transaction order is determined based on the prediction result.

[0074] Specifically, the target data includes: historical customer service record information related to the target commodity transaction order, browsing and clicking record information performed by the user in the current commodity information service system client before initiating the customer service request, and / or status information of the target commodity transaction order in multiple dimensions.

[0075] The historical customer service record information related to the target commodity transaction order includes: record information generated during the communication between the user and the merchant user, the manual customer service and / or the intelligent customer service on the system side regarding the target commodity transaction order.

[0076] Specifically, the large model generation unit can be used for:

[0077] The acquired target data is arranged and processed according to the time dimension to generate timeline trajectory information, and the timeline trajectory information is determined as input data for the AI ​​generation model.

[0078] In addition, the large model generation unit can also be used to:

[0079] In the process of arranging the acquired target data, key data related to problem prediction is screened out, and the timeline trajectory information is generated according to the screened out key data.

[0080] In addition, an embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of any one of the methods in the aforementioned method embodiments are implemented.

[0081] And an electronic device, comprising:

[0082] one or more processors; and

[0083] A memory associated with the one or more processors, the memory being used to store program instructions, wherein the program instructions, when read and executed by the one or more processors, execute the steps of the method described in any one of the aforementioned method embodiments.

[0084] A computer program product includes a computer program / computer executable instructions, which, when executed by a processor in an electronic device, implement the steps of the method described in the aforementioned method embodiment.

[0085] in, Figure 4The architecture of the electronic device is shown as an example, which may include a processor 410, a video display adapter 411, a disk drive 412, an input / output interface 413, a network interface 414, and a memory 420. The processor 410, the video display adapter 411, the disk drive 412, the input / output interface 413, the network interface 414, and the memory 420 may be communicatively connected via a communication bus 430.

[0086] Among them, the processor 410 can be implemented by a general-purpose CPU (Central Processing Unit, processor), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, etc., to execute relevant programs to implement the technical solution provided in this application.

[0087] The memory 420 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 420 can store an operating system 421 for controlling the operation of the electronic device 400, and a basic input and output system (BIOS) for controlling the low-level operation of the electronic device 400. In addition, a web browser 423, a data storage management system 424, and a problem prediction processing system 425, etc. can also be stored. The above-mentioned problem prediction processing system 425 can be an application program that specifically implements the operations of the aforementioned steps in the embodiment of the present application. In short, when the technical solution provided by the present application is implemented by software or firmware, the relevant program code is stored in the memory 420 and is called and executed by the processor 410.

[0088] The input / output interface 413 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0089] The network interface 414 is used to connect to a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0090] The bus 430 comprises a pathway for transmitting information between the various components of the device (eg, the processor 410, the video display adapter 411, the disk drive 412, the input / output interface 413, the network interface 414, and the memory 420).

[0091] It should be noted that, although the above device only shows a processor 410, a video display adapter 411, a disk drive 412, an input / output interface 413, a network interface 414, a memory 420, a bus 430, etc., in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include components necessary for implementing the solution of the present application, and does not necessarily include all the components shown in the figure.

[0092] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly contributed to the prior art in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

[0093] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0094] The problem prediction method and electronic device provided by the present application are introduced in detail above. The principle and implementation method of the present application are explained by using specific examples in this article. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. At the same time, for those skilled in the art, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as limiting the present application.

Claims

1. A problem prediction method, characterized in that: include: In response to a service request initiated by a user, determining an associated target commodity transaction order; Acquire target data related to problem classification prediction of the target commodity transaction order; Determine input data of an artificial intelligence (AI) generation model according to the acquired target data, so that the AI ​​generation model generates a user's demand text or a question text to be consulted based on the target data; The user demand text or question text generated by the AI ​​generation model is input into a preset classification prediction model to be converted into a standardized question description text in a candidate set, and the standardized question description text is provided to the client for user selection.

2. The method according to claim 1, characterized in that The determining of the associated target commodity transaction order includes: If the user initiates a customer service request after selecting an order, the order selected by the user is determined as the target commodity transaction order.

3. The method according to claim 1, characterized in that The determining of the associated target commodity transaction order includes: If the user has not selected an order before initiating a customer service request, a prediction of the problematic order is performed based on the status of each order in the order list associated with the user, and the target commodity transaction order is determined based on the prediction result.

4. The method according to claim 1, characterized in that: The target data includes: historical customer service record information related to the target commodity transaction order, browsing and clicking record information performed by the user in the current commodity information service system client before initiating the customer service request, and / or status information of the target commodity transaction order in multiple dimensions.

5. The method according to claim 4, characterized in that The historical customer service record information related to the target commodity transaction order includes: record information generated during the communication between the user and the merchant user, the manual customer service and / or the intelligent customer service on the system side regarding the target commodity transaction order.

6. The method according to claim 1, characterized in that The step of determining the input data of the artificial intelligence AI generation model according to the acquired target data includes: The acquired target data is arranged and processed according to the time dimension to generate timeline trajectory information, and the timeline trajectory information is determined as input data for the AI ​​generation model.

7. The method according to claim 6, characterized in that Also includes: In the process of arranging the acquired target data, key data related to problem prediction is screened out, and the timeline trajectory information is generated according to the screened out key data.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 7 are implemented.

9. An electronic device, characterized in that: include: one or more processors; as well as A memory associated with the one or more processors, the memory being used to store program instructions, wherein the program instructions, when read and executed by the one or more processors, execute the steps of the method described in any one of claims 1 to 7.

10. A computer program product comprising a computer program / computer executable instructions, characterized in that: When the computer program / computer executable instructions are executed by a processor in an electronic device, the steps of the method according to any one of claims 1 to 7 are implemented.

Images