After-sales service method, device, electronic device and storage medium

Abstract

The present application provides a method, device, electronic device and storage medium for after-sales service, wherein the after-sales service method comprises: obtaining evaluation information of a product; when it is determined that the evaluation information is a negative evaluation, determining a first meta-event of the evaluation information; based on the first meta-event and a pre-stored causal graph, determining the target reason for making the evaluation information on the product, wherein the causal graph includes the reason for the first meta-event; and outputting the target reason so that the user can perform after-sales service based on the target reason.

Description

Technical Field

[0001] The present application relates to the technical field of e-commerce, and in particular to a method, device, electronic device and storage medium for after-sales service. Background Art

[0002] As e-commerce gradually enters people's lives, online shopping has become a common way for people to shop. After shopping online, people usually evaluate the products. Merchants will conduct after-sales visits to buyers who have negative comments on the products. The usual practice is to manually select negative comments or use sentiment analysis to select corresponding buyers, conduct after-sales visits one by one, track their shortcomings of the products and make records to support future optimization of the products. However, traditional sentiment analysis can only distinguish relatively obvious product reviews, and it is difficult to parse the sentiment bias of ironic, rhetorical, and metaphorical sentences. These negative comments cannot be identified, resulting in data omissions. In addition, after extracting the seller's evaluation, the relationship between the evaluation and the entity is displayed through the knowledge graph to facilitate customer service. However, the relationship in the knowledge graph is relatively fixed and static knowledge, and many comments are often logical. For example, when the washing machine is working, the sound is a bit loud. When extracting the knowledge graph, it can only be proposed that the washing machine-working->loud sound. However, the loud noise of the washing machine is often caused by the imbalance of the balancing liquid in the washing tub during the rolling process. This part of the information cannot be seen in the knowledge graph, resulting in the customer service cannot understand the user's true expression, and the after-sales service efficiency is low. Summary of the invention

[0003] In response to the above problems, the present application provides a method, device, electronic device and storage medium for after-sales service.

[0004] This application provides an after-sales service method, including:

[0005] Get product evaluation information;

[0006] In the case where it is determined that the evaluation information is a negative evaluation, determining a first meta-event of the evaluation information;

[0007] Determining a target reason for making the evaluation information on the product based on the first meta-event and a pre-stored cause-and-effect graph, wherein the cause of the first meta-event is included in the cause-and-effect graph;

[0008] The target reason is outputted so that the user can perform after-sales service based on the target reason.

[0009] In some embodiments, the method further comprises:

[0010] Inputting the evaluation information into a first sentiment analysis model to determine a first probability that the evaluation information is a negative evaluation, wherein the first sentiment analysis model is capable of identifying the probability that the evaluation information with explicit sentiment is a negative evaluation;

[0011] Based on the first probability and a first preset probability threshold, it is determined whether the evaluation information is a negative evaluation.

[0012] In some embodiments, the method further comprises:

[0013] In the case where it is determined that the evaluation information is not a negative evaluation, determining that the evaluation information is suspected negative evaluation information;

[0014] Inputting the suspected negative evaluation information into a second sentiment analysis model to determine a second probability that the suspected negative evaluation is a negative evaluation, wherein the second sentiment analysis model is capable of identifying the probability that evaluation information with implicit sentiment is a negative evaluation;

[0015] It is determined whether the suspected evaluation information is a negative evaluation based on the second probability and a second preset probability threshold.

[0016] In some embodiments, the method further comprises:

[0017] Acquire a sample corpus, wherein the sample corpus is annotated with implicit sentiment words and the context of the sample corpus displays sentiment classification;

[0018] Based on the sample corpus, determining the implicit sentiment words, the contextual display sentiment classification and the text vector of the sample corpus based on the semantic relationship;

[0019] Determining a second sentiment analysis model based on the implicit sentiment words, the contextual display sentiment classification and the text vector;

[0020] The second sentiment analysis model is stored.

[0021] In some embodiments, the determining of the second sentiment analysis model based on the implicit sentiment words, the context display sentiment classification and the text vector comprises:

[0022] Determining a first neural network model based on the implicit sentiment words;

[0023] Determining a second neural network model based on the context display sentiment classification;

[0024] Determine a third neural network model based on the text vector;

[0025] The first neural network model, the second neural network model and the third neural network model are combined to obtain a second sentiment analysis model.

[0026] In some embodiments, the method further comprises:

[0027] Determine a second meta-event based on the target cause and the event graph;

[0028] The second meta-event is outputted so that the user can perform after-sales service based on the target reason and the second meta-event.

[0029] In some embodiments, determining the target reason for making the evaluation information on the product based on the first meta-event information and a pre-stored cause-and-effect graph includes:

[0030] Determine the target word of the first meta-event;

[0031] Matching the target word with words in a pre-stored event graph to determine a matching result;

[0032] Based on the matching result, a target reason for making the evaluation information on the product is determined.

[0033] The present application provides an after-sales service device, including:

[0034] The first acquisition module is used to obtain the evaluation information of the product;

[0035] A first determining module, configured to determine a first meta-event of the evaluation information when it is determined that the evaluation information is a negative evaluation;

[0036] A second determination module is used to determine the target reason for making the evaluation information on the product based on the first meta-event and a pre-stored cause-and-effect graph, wherein the cause of the first meta-event is included in the cause-and-effect graph;

[0037] The first output module is used to output the target reason so that the user can perform after-sales service based on the target reason.

[0038] An embodiment of the present application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the after-sales service method described above is executed.

[0039] An embodiment of the present application provides a storage medium, in which computer-executable instructions are stored. The computer-executable instructions are used to execute any of the after-sales service methods described above.

[0040] The present application provides an after-sales service method, device, electronic device and storage medium. When it is determined that the evaluation information is a negative evaluation, the first meta-event of the evaluation information is determined; based on the first meta-event and a pre-stored causal map, the target reason for making the evaluation information on the product is determined; the target reason is output, so that the customer service can intuitively understand the cause of the first meta-event and understand the true expression of the user in making the evaluation, thereby improving the efficiency of after-sales service. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Hereinafter, the present application will be described in more detail based on embodiments and with reference to the accompanying drawings.

[0042] Figure 1 A schematic diagram of an implementation flow of an after-sales service method provided in an embodiment of the present application;

[0043] Figure 2 A schematic diagram of an implementation flow of establishing a second sentiment analysis model provided in an embodiment of the present application;

[0044] Figure 3 A schematic diagram of an implementation flow of an after-sales service method provided in an embodiment of the present application;

[0045] Figure 4 A schematic diagram of a process for establishing an implicit sentiment analysis model provided in an embodiment of the present application;

[0046] Figure 5 A schematic diagram of the structure of a washing machine event map provided in an embodiment of the present application;

[0047] Figure 6 A schematic diagram of the structure of an after-sales service device provided in an embodiment of the present application;

[0048] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0049] In the drawings, the same reference numerals are used for the same components and the drawings are not drawn to scale. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0051] In the following description, reference is made to “some embodiments”, which describe a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0052] If similar descriptions of "first\second\third" appear in the application documents, the following instructions will be added. In the following description, the terms "first\second\third" involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It can be understood that "first\second\third" can be interchanged in a specific order or sequence where permitted, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0053] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0054] The embodiment of the present application provides a method for after-sales service, and the method is applied to an electronic device. The electronic device may be a mobile terminal, a computer, etc. The function implemented by the method for after-sales service provided by the embodiment of the present application can be implemented by calling a program code by a processor of the electronic device, wherein the program code can be stored in a computer storage medium. The embodiment of the present application provides a method for after-sales service, Figure 1 A schematic diagram of the implementation flow of an after-sales service method provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, including:

[0055] Step S101, obtaining evaluation information of the product.

[0056] In an embodiment of the present application, the commodity may be a washing machine, a television, an air conditioner, a mobile terminal, a computer, etc., and the evaluation information is the evaluation of the commodity by the user after purchasing the commodity. The evaluation information may include an evaluation showing emotions. For example, the commodity is a washing machine, and the evaluation showing emotions is: This washing machine is really good. The evaluation showing emotions is a direct expression of user feelings by the user. In some embodiments, the evaluation information may be an evaluation of implicit emotions. In the evaluation of implicit emotions, the sentence itself does not contain explicit emotional words, but expresses subjective emotions. The evaluation of implicit emotions is usually divided into factual implicit emotions and rhetorical implicit emotions. Rhetorical metaphor, metaphor, rhetorical question and irony. For example, the evaluation of implicit emotions can be that this washing machine is like a little helper doing housework, this washing machine is a bit silly, where is the intelligence?

[0057] Step S102: when it is determined that the evaluation information is a negative evaluation, determine a first meta-event of the evaluation information.

[0058] In the embodiment of the present application, before step S102, it is also necessary to determine whether the evaluation information is a positive evaluation or a negative evaluation. The first probability of the evaluation information being a negative evaluation can be determined by the first sentiment analysis model. The first sentiment analysis model can identify the probability that the evaluation information with explicit emotions is a negative evaluation; based on the first probability and the first preset probability threshold, determine whether the evaluation information is a negative evaluation. In the embodiment of the present application, when the first probability is greater than the first preset probability threshold, the evaluation information is determined to be a negative evaluation, and when the first probability is less than the first preset probability threshold, it is determined that the evaluation information is suspected negative evaluation information; the suspected negative evaluation information is input into the second sentiment analysis model to determine the second probability that the suspected negative evaluation is a negative evaluation, and the second sentiment analysis model can identify the probability that the evaluation information with implicit emotions is a negative evaluation; based on the second probability and the second preset probability threshold, determine whether the suspected evaluation information is a negative evaluation. In the embodiment of the present application, when the second probability is greater than the second probability threshold, determine that the suspected evaluation information is a negative evaluation, and when the second probability is less than the second probability threshold, determine that the suspected negative evaluation information is not a negative evaluation.

[0059] In the embodiment of the present application, the evaluation information can be subjected to meta-event extraction to obtain the first meta-event of the evaluation information. In the embodiment of the present application, the first meta-event can be multiple or one. For example, if the evaluation information is that the washing machine makes a loud sound when it is working, then the first meta-event is: the washing machine makes a loud sound when it is working.

[0060] Step S103: determining the target reason for making the evaluation information on the product based on the first meta-event and the pre-stored cause-and-effect graph.

[0061] In the embodiment of the present application, the causal graph is pre-stored, and the causal graph includes the cause of the first meta-event. Based on the first meta-event and the pre-stored causal graph, the target cause of the evaluation information on the product can be determined. When it is implemented, the target word of the first meta-event can be determined; the target word is matched with the words in the pre-stored causal graph to determine the matching result; based on the matching result, the target cause of the evaluation information on the product can be determined. Continuing from the above, the causal event of the loud noise of the washing machine can be determined through the causal graph. There can be many causal events. For example, an imbalance of the balancing fluid causes the transmitter to make too loud noise, or a loose screw causes the noise to be too loud. In the embodiment of the present application, the target cause can be: imbalance of the balancing fluid, loose screws, etc.

[0062] Step S104: outputting the target reason so that the user can perform after-sales service based on the target reason.

[0063] In the embodiment of the present application, after the target reason is determined, the target reason can be output, and the customer service can provide after-sales service based on the target reason. The after-sales service here can be a return visit to multiple users. After locating the specific problem, it can be fed back to the after-sales maintenance department to find the corresponding maintenance personnel for on-site maintenance. In the embodiment of the present application, after the target reason is output, the customer can quickly understand the user's problem and the mechanical failure, avoiding a poor user experience due to the customer service department's unfamiliarity with the user's communication.

[0064] The present application provides an after-sales service method, which, when it is determined that the evaluation information is a negative evaluation, determines the first meta-event of the evaluation information; based on the first meta-event and a pre-stored causal graph, determines the target reason for making the evaluation information on the product; outputs the target reason, which enables the customer service staff to intuitively understand the cause of the first meta-event and understand the true expression of the user in making the evaluation, thereby improving the efficiency of after-sales service.

[0065] In some embodiments, before step S102 "when determining that the evaluation information is a negative evaluation, determining a first meta-event of the evaluation information", the method further includes:

[0066] Step S1, inputting the evaluation information into a first sentiment analysis model to determine a first probability that the evaluation information is a negative evaluation, wherein the first sentiment analysis model can identify the probability that the evaluation information with explicit sentiment is a negative evaluation.

[0067] In the embodiment of the present application, the first sentiment analysis model is pre-trained, and the first sentiment analysis model is obtained by training a neural network model. In the embodiment of the present application, samples with probabilities of displaying emotions and evaluation types are input into the neural network model for training, thereby obtaining evaluation types that can identify evaluation information that displays emotions.

[0068] Step S2: determining whether the evaluation information is a negative evaluation based on the first probability and a first preset probability threshold.

[0069] The first probability can be compared with a first preset probability threshold. For example, when the first probability is greater than the first threshold probability threshold, the evaluation information can be determined to be a negative evaluation. When the first probability is less than the first preset probability threshold, the evaluation information can be determined to be a positive evaluation. In the embodiment of the present application, a negative evaluation is a bad evaluation, and a positive evaluation is a good evaluation.

[0070] In some embodiments, after step S2 of "determining whether the evaluation information is a negative evaluation based on the first probability and a first preset probability threshold", the method further includes:

[0071] Step S3: when it is determined that the evaluation information is not a negative evaluation, determine that the evaluation information is suspected negative evaluation information.

[0072] In the embodiment of the present application, since the first sentiment analysis model can identify evaluations that show emotions, but cannot identify evaluations that show implicit emotions, when the first sentiment analysis model determines that the evaluation is not negative, it is possible that the user used negative evaluations with implicit emotions. Therefore, the evaluation information is determined as suspected negative evaluation information here.

[0073] Step S4: input the suspected negative evaluation information into a second sentiment analysis model to determine a second probability that the suspected negative evaluation is a negative evaluation.

[0074] In the embodiment of the present application, the second sentiment analysis model can identify the probability that the evaluation information with implicit sentiment is a negative evaluation. In the embodiment of the present application, the implicit sentiment itself does not contain explicit sentiment words, but expresses subjective sentiment. Implicit sentiment sentences are usually divided into factual implicit sentiment and rhetorical implicit sentiment. Rhetorical types include metaphorical, metaphorical, rhetorical and ironic.

[0075] In an embodiment of the present application, the second sentiment analysis model is pre-established and can be obtained by acquiring a sample corpus, wherein the sample corpus is annotated with implicit sentiment words and the context display sentiment classification of the sample corpus; based on the sample corpus, determining the implicit sentiment words, the context display sentiment classification and the text vector of the sample corpus based on the semantic relationship; determining the second sentiment analysis model based on the implicit sentiment words, the context display sentiment classification and the text vector; thereby storing the second sentiment analysis model.

[0076] Step S5: determining whether the suspected evaluation information is a negative evaluation based on the second probability and a second preset probability threshold.

[0077] In the embodiment of the present application, when the second probability is greater than the second preset probability threshold, the suspected negative evaluation information is determined to be a negative evaluation. When the second probability is less than the second threshold probability threshold, the suspected negative evaluation information is not a negative evaluation. When it is not a negative evaluation, the evaluation information can be a neutral evaluation or a positive evaluation.

[0078] Before step S4 "inputting the suspected negative evaluation information into the second sentiment analysis model to determine the second probability that the suspected negative evaluation is a negative evaluation", the method further includes establishing a second sentiment analysis model, Figure 2 A schematic diagram of an implementation flow of establishing a second sentiment analysis model provided in an embodiment of the present application is as follows: Figure 2 As shown, including:

[0079] Step S11, obtaining sample corpus.

[0080] In the embodiment of the present application, the sample corpus is annotated with implicit sentiment words and the context of the sample corpus displays sentiment classification.

[0081] Step S12, based on the sample corpus, determining the implicit sentiment words, the context display sentiment classification and the text vector of the sample corpus based on the semantic relationship;

[0082] Step S13, determining a second sentiment analysis model based on the implicit sentiment words, the context display sentiment classification and the text vector.

[0083] In an embodiment of the present application, a first neural network model can be determined based on the implicit sentiment words; a second neural network model can be determined based on the context display sentiment classification; a third neural network model can be determined based on the text vector; and the first neural network model, the second neural network model and the third neural network model can be merged to obtain a second sentiment analysis model.

[0084] Step S14: storing the second sentiment analysis model.

[0085] In some embodiments, after obtaining the target reason, the following steps may also be included:

[0086] Step S105, determining a second meta-event based on the target cause and the event graph.

[0087] In an embodiment of the present application, the second meta-event is different from the first meta-event, and there is a causal relationship between the second meta-event and the target cause. For example, the problem of the washing machine balance liquid may have other manifestations, and the other manifestations are stored in the event map, thereby determining the second meta-event.

[0088] Step S106: output the second meta-event, so that the user can perform after-sales service based on the target reason and the second meta-event.

[0089] In the embodiment of the present application, customer service can provide better after-sales service through the target reason and the second meta-event.

[0090] In some embodiments, step S103 "determining the target reason for making the evaluation information on the product based on the first meta-event information and the pre-stored cause-and-effect graph" can be implemented by the following steps:

[0091] Step S1031, determining the target word of the first meta-event;

[0092] Step S1032, matching the target word with words in a pre-stored event graph to determine a matching result;

[0093] Step S1033: determining the target reason for making the evaluation information on the product based on the matching result.

[0094] The present application embodiment provides a method for after-sales service. Figure 3 A schematic diagram of the implementation flow of an after-sales service method provided in an embodiment of the present application is shown in FIG. Figure 3 As shown, including:

[0095] Step S21, obtaining comments on the product (same as the evaluation information in the above embodiment).

[0096] Step S22: sentiment analysis.

[0097] In the embodiment of the present application, sentiment analysis is performed through a first sentiment analysis model.

[0098] Step S23, determining whether the comment is a negative evaluation.

[0099] In the embodiment of the present application, a first probability is obtained after sentiment analysis, and whether the comment is a negative evaluation can be determined based on the first probability and a first preset probability threshold.

[0100] In the embodiment of the present application, when the comment is a negative evaluation, step S24 is executed, and when the comment is not a negative evaluation, step S25 is executed.

[0101] Step S24, storing the event graph.

[0102] In the embodiment of the present application, the cause-and-effect relationship between the first element event and the cause is stored in the event graph.

[0103] Step S25, implicit sentiment analysis.

[0104] In the embodiment of the present application, implicit sentiment analysis is performed by a second sentiment analysis model. In an implicit sentiment sentence, the sentence itself does not contain explicit sentiment words, but expresses subjective sentiment. Implicit sentiment sentences are usually divided into factual implicit sentiment and rhetorical implicit sentiment. Rhetorical metaphor, metaphor, rhetorical question and irony. In the present invention, the main focus is on classifying the sentiment tendency of implicit sentiment sentences. The following are examples of explicit and implicit sentiments:

[0105]

[0106] Step S26, determining whether the comment is a negative evaluation.

[0107] In the embodiment of the present application, when the comment is a negative comment, step S24 is executed. When the comment is not a negative comment, the process ends.

[0108] Step S27, storing the event graph.

[0109] Step S28, extracting causal events in the comments.

[0110] Step S29: performing comment analysis on multiple information streams.

[0111] Step S30, outputting the comment analysis result (same as the target reason in the above embodiment).

[0112] Step S31, push the analysis results to enable customer service to make after-sales return visits.

[0113] In the embodiment of the present application, before step S25, the method further includes: establishing an implicit sentiment analysis model (same as the second sentiment analysis model in the above embodiment), and the establishment process can be implemented by the following steps: Figure 4 A schematic diagram of a process for establishing an implicit sentiment analysis model provided in an embodiment of the present application is shown in FIG. Figure 4 As shown, including:

[0114] Step S41, obtaining corpus marked with sentiment tendency (same as the sample corpus in the above embodiment).

[0115] Step S42, determining the implicit sentiment words of the corpus, the context-displayed sentiment classification, and the multi-layer convolutional neural network text vector based on syntactic and semantic relations.

[0116] Step S43, performing multi-dimensional merging.

[0117] Step S44, constructing a model (the same as the second sentiment analysis model in the above embodiment).

[0118] Step S45, output the model.

[0119] In this way, the implicit sentiment analysis model is established, and the model building process is completed.

[0120] In the embodiment of the present application, step S24 can be implemented by the following steps:

[0121] Step S51, extract meta-events from comments to obtain multiple original meta-events. For example, when a washing machine is working, the sound is very loud. Then the meta-event extracted is washing machine--working-->loud sound.

[0122] Step S52, matching is performed based on a preset event graph and meta-events, where the event graph includes events connected according to a causal relationship or an event time sequence relationship.

[0123] Step S53, obtaining the causal relationship of the meta-event by matching with the preset event graph.

[0124] Figure 5 A schematic diagram of the structure of a washing machine event map provided in an embodiment of the present application is shown in FIG. Figure 5 As shown,

[0125] pass Figure 5 Through the principle diagram, it is determined that the loud noise of the washing machine is due to engine problems, or the loud noise of the washing machine is due to balance fluid problems.

[0126] The after-sales service method provided in the embodiment of the present application, through implicit sentiment analysis, parses out ironic, rhetorical, and metaphorical product reviews that cannot be identified by traditional sentiment analysis, collects more product reviews from buyers, and analyzes suggestions or product deficiencies that are conducive to product improvement and optimization. The atlas of events depicts a logical society, and the research object is predicate events and their internal and external connections. The user's comments are logically sorted out to understand the user's expression in a timely manner, and the customer service can sort out the user's true expression to improve the customer service's after-sales service efficiency.

[0127] Based on the foregoing embodiments, the embodiments of the present application provide an after-sales service device, and the modules included in the device and the units included in the modules can be implemented by a processor in a computer device; of course, they can also be implemented by a specific logic circuit; in the implementation process, the processor can be a central processing unit (CPU, Central Processing Unit), a microprocessor (MPU, Microprocessor Unit), a digital signal processor (DSP, Digital Signal Processing) or a field programmable gate array (FPGA, Field Programmable Gate Array), etc.

[0128] The present application embodiment provides an after-sales service device. Figure 6 This is a schematic diagram of the structure of an after-sales service device provided in an embodiment of the present application. The after-sales service device 600 includes:

[0129] The first acquisition module 601 is used to acquire evaluation information of the product;

[0130] A first determining module 602, configured to determine a first meta-event of the evaluation information when it is determined that the evaluation information is a negative evaluation;

[0131] A second determination module 603 is used to determine the target reason for making the evaluation information on the product based on the first meta-event and a pre-stored cause-and-effect graph, wherein the cause of the first meta-event is included in the cause-and-effect graph;

[0132] The first output module 604 is used to output the target reason so that the user can perform after-sales service based on the target reason.

[0133] In some embodiments, the after-sales service device 600 further includes:

[0134] A third determination module is used to input the evaluation information into a first sentiment analysis model to determine a first probability that the evaluation information is a negative evaluation, wherein the first sentiment analysis model can identify the probability that the evaluation information with explicit sentiment is a negative evaluation;

[0135] The fourth determination module is used to determine whether the evaluation information is a negative evaluation based on the first probability and a first preset probability threshold.

[0136] In some embodiments, the after-sales service device 600 further includes:

[0137] a fifth determining module, configured to determine that the evaluation information is suspected negative evaluation information when it is determined that the evaluation information is not a negative evaluation;

[0138] a sixth determination module, configured to input the suspected negative evaluation information into a second sentiment analysis model to determine a second probability that the suspected negative evaluation is a negative evaluation, wherein the second sentiment analysis model is capable of identifying the probability that evaluation information with implicit sentiment is a negative evaluation;

[0139] It is determined whether the suspected negative evaluation information is a negative evaluation based on the second probability and a second preset probability threshold.

[0140] In some embodiments, the after-sales service device 600 further includes:

[0141] A second acquisition module is used to acquire sample corpus, wherein the sample corpus is annotated with implicit sentiment words and the context of the sample corpus shows sentiment classification;

[0142] A seventh determination module, for determining, based on the sample corpus, the implicit sentiment words, the context display sentiment classification and the text vector of the sample corpus based on semantic relations;

[0143] An eighth determination module, configured to determine a second sentiment analysis model based on the implicit sentiment words, the context display sentiment classification and the text vector;

[0144] The second sentiment analysis model is stored.

[0145] In some embodiments, the seventh determination module includes:

[0146] A first determining unit, configured to determine a first neural network model based on the implicit sentiment words;

[0147] A second determining unit, configured to determine a second neural network model based on the context display sentiment classification;

[0148] A third determining unit, configured to determine a third neural network model based on the text vector;

[0149] A merging unit is used to merge the first neural network model, the second neural network model and the third neural network model to obtain a second sentiment analysis model.

[0150] In some embodiments, the after-sales service device 600 further includes:

[0151] A ninth determination module, configured to determine a second meta-event based on the target cause and the event graph;

[0152] The second output module is used to output the second meta-event so that the user can perform after-sales service based on the target reason and the second meta-event.

[0153] In some embodiments, the second determining module 603 includes:

[0154] A fourth determining unit, used for determining a target word of the first meta-event;

[0155] A fifth determination unit, used to match the target word with words in a pre-stored event graph to determine a matching result;

[0156] The sixth determining unit is used to determine the target reason for making the evaluation information on the product based on the matching result.

[0157] It should be noted that in the embodiments of the present application, if the above-mentioned after-sales service method is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM, Read Only Memory), a disk or an optical disk. In this way, the embodiments of the present application are not limited to any specific combination of hardware and software.

[0158] Accordingly, an embodiment of the present application provides a storage medium on which a computer program is stored, characterized in that when the computer program is executed by a processor, the steps in the after-sales service method provided in the above embodiment are implemented.

[0159] An embodiment of the present application provides an electronic device; Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown in FIG. Figure 7 As shown, the electronic device 700 includes: a processor 701, at least one communication bus 702, a user interface 703, at least one external communication interface 704, and a memory 705. The communication bus 702 is configured to realize the connection and communication between these components. The user interface 703 may include a display screen, and the external communication interface 704 may include a standard wired interface and a wireless interface. The processor 701 is configured to execute the program of the after-sales service method stored in the memory to implement the steps in the after-sales service method provided in the above embodiment.

[0160] The description of the above display device and storage medium embodiments is similar to the description of the above method embodiments, and has similar beneficial effects as the method embodiments. For technical details not disclosed in the computer device and storage medium embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0161] It should be noted here that the description of the above storage medium and device embodiments is similar to the description of the above method embodiments, and has similar beneficial effects as the method embodiments. For technical details not disclosed in the storage medium and device embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0162] It should be understood that "one embodiment" or "an embodiment" mentioned throughout the specification means that specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. It should be understood that in various embodiments of the present application, the size of the sequence number of the above-mentioned processes does not mean the order of execution, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application. The above-mentioned sequence numbers of the embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments.

[0163] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0164] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0165] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0166] In addition, all functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may be a separate unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0167] A person skilled in the art can understand that all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read Only Memory), disks or optical disks, etc. Various media that can store program codes.

[0168] Alternatively, if the above-mentioned integrated unit of the present application is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application can essentially or in other words, the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for a controller to execute all or part of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0169] The above is only an implementation method of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for after-sales service, characterized in that: include: Get product evaluation information; Inputting the evaluation information into a first sentiment analysis model to determine a first probability that the evaluation information is a negative evaluation, wherein the first sentiment analysis model is capable of identifying the probability that the evaluation information with explicit sentiment is a negative evaluation; Based on the first probability and a first preset probability threshold, determining whether the evaluation information is a negative evaluation: when the first probability is greater than the first preset probability threshold, determining that the evaluation information is a negative evaluation; when the first probability is less than the first preset probability threshold, determining that the evaluation information is suspected negative evaluation information; In the case where it is determined that the evaluation information is not a negative evaluation, determining that the evaluation information is suspected negative evaluation information; Inputting the suspected negative evaluation information into a second sentiment analysis model to determine a second probability that the suspected negative evaluation information is a negative evaluation, wherein the second sentiment analysis model is capable of identifying the probability that evaluation information with implicit sentiment is a negative evaluation; Determine whether the suspected negative evaluation information is a negative evaluation based on the second probability and a second preset probability threshold: when the second probability is greater than the second probability threshold, determine that the suspected negative evaluation information is a negative evaluation; when the second probability is less than the second probability threshold, determine that the suspected negative evaluation information is not a negative evaluation; When it is determined that the evaluation information is a negative evaluation, extracting meta-events from the evaluation information to determine a first meta-event of the evaluation information; Based on the first meta-event and the pre-stored cause-and-effect graph, determining the target reason for making the evaluation information on the product, including: determining the target word of the first meta-event; matching the target word with the words in the pre-stored cause-and-effect graph to determine the matching result; determining the target reason for making the evaluation information on the product based on the matching result, wherein the cause of the first meta-event is included in the cause-and-effect graph, and the cause-and-effect graph also includes events connected according to a causal relationship, wherein the events connected according to the causal relationship include: after-sales maintenance reasons; The target reason is outputted so that the user can perform after-sales service based on the target reason.

2. The method according to claim 1, characterized in that The method further comprises: Acquire a sample corpus, wherein the sample corpus is annotated with implicit sentiment words and the context of the sample corpus displays sentiment classification; Based on the sample corpus, determining the implicit sentiment words, the contextual display sentiment classification and the text vector of the sample corpus based on the semantic relationship; Determining a second sentiment analysis model based on the implicit sentiment words, the contextual display sentiment classification and the text vector; The second sentiment analysis model is stored.

3. The method according to claim 2, characterized in that The determining of the second sentiment analysis model based on the implicit sentiment words, the context display sentiment classification and the text vector comprises: Determining a first neural network model based on the implicit sentiment words; Determining a second neural network model based on the context display sentiment classification; Determine a third neural network model based on the text vector; The first neural network model, the second neural network model and the third neural network model are combined to obtain a second sentiment analysis model.

4. The method according to claim 1, characterized in that: The method further comprises: Determine a second meta-event based on the target cause and the event graph; The second meta-event is outputted so that the user can perform after-sales service based on the target reason and the second meta-event.

5. An after-sales service device, characterized in that: include: The first acquisition module is used to obtain the evaluation information of the product; A first determination module is used to input the evaluation information into a first sentiment analysis model to determine a first probability that the evaluation information is a negative evaluation, wherein the first sentiment analysis model can identify the probability that the evaluation information with explicit sentiment is a negative evaluation; based on the first probability and a first preset probability threshold, determine whether the evaluation information is a negative evaluation: when the first probability is greater than the first preset probability threshold, determine that the evaluation information is a negative evaluation; when the first probability is less than the first preset probability threshold, determine that the evaluation information is suspected negative evaluation information; when it is determined that the evaluation information is not a negative evaluation, determine that the evaluation information is suspected negative evaluation information; Input the suspected negative evaluation information into a second sentiment analysis model to determine a second probability that the suspected negative evaluation information is a negative evaluation, wherein the second sentiment analysis model can identify the probability that evaluation information with implicit emotions is a negative evaluation; determine whether the suspected negative evaluation information is a negative evaluation based on the second probability and a second preset probability threshold: when the second probability is greater than the second probability threshold, determine that the suspected negative evaluation information is a negative evaluation; when the second probability is less than the second probability threshold, determine that the suspected negative evaluation information is not a negative evaluation; When it is determined that the evaluation information is a negative evaluation, extracting meta-events from the evaluation information to determine a first meta-event of the evaluation information; The second determination module is used to determine the target reason for making the evaluation information on the product based on the first meta-event and the pre-stored cause-and-effect graph, including: determining the target word of the first meta-event; matching the target word with the words in the pre-stored cause-and-effect graph to determine the matching result; determining the target reason for making the evaluation information on the product based on the matching result, wherein the cause of the first meta-event is included in the cause-and-effect graph, and the cause-and-effect graph also includes events connected according to a causal relationship, wherein the events connected according to the causal relationship include: after-sales maintenance reasons; The first output module is used to output the target reason so that the user can perform after-sales service based on the target reason.

6. An electronic device, characterized in that: The method comprises a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the after-sales service method according to any one of claims 1 to 4 is executed.

7. A storage medium, characterized in that: The storage medium stores computer executable instructions, and the computer executable instructions are used to execute the after-sales service method according to any one of claims 1 to 4.

Images