Method and system for intelligently providing after-sales service based on big data

Abstract

The invention belongs to the technical field of after-sales service. The invention provides an after-sales service intelligence providing method and system based on big data. The method comprises the following steps: acquiring after-sales associated big data related to a target brand, and constructing an after-sales knowledge data set based on the after-sales associated big data; receiving first consultation information input by a target client, and determining a question element extraction strategy based on the first consultation information; receiving second consultation information input by a target client, extracting a plurality of second question elements from the second consultation information, constructing a target question element based on each second question element and the question element extraction strategy, and performing matching from the after-sales knowledge data set based on the target question element to obtain target after-sales knowledge data; and outputting the target after-sales knowledge data to the target customer. According to the invention, the'thinking 'time of the intelligent customer service robot can be effectively shortened, and better after-sales service experience is provided for customers.

Description

Technical Field

[0001] The present invention relates to the field of after-sales service technology, and in particular to a method and system for intelligently providing after-sales service based on big data. Background Art

[0002] Online after-sales service generally refers to customer service and support provided through the Internet, which includes but is not limited to online customer service, remote technical support, self-service portal, email support and social media interaction, etc. This service model allows customers to get help through different online channels anytime and anywhere without having to go to a physical service point.

[0003] Nowadays, more and more online after-sales services are equipped with intelligent customer service robots, which can understand and respond to customer inquiries and provide automated customer service. However, in order to ensure the accuracy of the answers, the intelligent customer service robots search and respond to the answers according to the full information elements contained in the questions entered by the customers, which causes the intelligent customer service robots to "think" for too long, and customers need to wait for a long time to get the answers from the intelligent customer service robots. The above reasons have led to the fact that the customer experience when interacting with the intelligent customer service robots on after-sales issues is still not good enough, and further improvement is needed. Summary of the invention

[0004] In this regard, the present invention provides a method, system, electronic device, computer storage medium and computer program product for intelligent provision of after-sales service based on big data to solve the above-mentioned technical problems.

[0005] The present invention provides a method for intelligently providing after-sales service based on big data, which is applied to an intelligent customer service robot and comprises the following steps:

[0006] Acquire after-sales related big data related to the target brand, and construct a after-sales knowledge dataset based on the after-sales related big data;

[0007] Receiving first consulting information input by a target customer, and determining a question element extraction strategy based on the first consulting information;

[0008] receiving second consulting information input by a target customer, extracting a plurality of second question elements from the second consulting information, constructing a target question element based on each of the second question elements and the question element extraction strategy, and matching the target after-sales knowledge data from the after-sales knowledge data set based on the target question element;

[0009] The target after-sales knowledge data is output to the target customer.

[0010] Optionally, the acquiring of after-sales related big data related to the target brand and constructing a after-sales knowledge dataset based on the after-sales related big data includes:

[0011] Determine product sequence information of the target brand, wherein the product sequence information includes a plurality of products arranged in a multi-level tree structure;

[0012] Using web crawler technology to crawl from the Internet to obtain after-sales related data corresponding to each of the products, each of the after-sales related data constitutes after-sales related big data;

[0013] Corresponding after-sales problems and corresponding solutions are extracted from each of the after-sales related data, and the after-sales problems are associated with the corresponding solutions to form an after-sales knowledge data. All the after-sales knowledge data are deduplicated, classified and integrated to obtain the after-sales knowledge data set.

[0014] Optionally, the crawling of after-sales related data corresponding to each of the products from the Internet using a web crawler technology includes:

[0015] Accessing an online shopping website, and extracting from the online shopping website a first set of product detail pages of a target brand and a second set of product detail pages of several competing brands of the target brand;

[0016] Performing semantic analysis on the first set of product detail pages to extract a first number of promotional elements; and performing semantic analysis on the second set of product detail pages to extract a second number of promotional elements, calculating an absolute value of a difference between the first number and the second number, and determining a third number based on the absolute value of the difference;

[0017] Based on the third quantity, after-sales related data corresponding to each of the products is crawled from the Internet using web crawler technology.

[0018] Optionally, the receiving first consultation information input by the target customer and determining a question element extraction strategy based on the first consultation information includes:

[0019] Displaying a plurality of question type buttons to the target customer, the buttons including a product category button and a plurality of product model buttons belonging to the product category button, receiving button selection information from the target customer, and determining the button selection information as the first consultation information of the target customer;

[0020] If the first consulting information corresponds to the product category button, the question element extraction strategy is set to a general strategy; if the first consulting information corresponds to the product model button, the question element extraction strategy is set to a precise strategy.

[0021] Optionally, extracting a plurality of second question elements from the second consulting information, and constructing a target question element based on each of the second question elements and the question element extraction strategy, includes:

[0022] Extracting a number of second question elements from the second consulting information;

[0023] If the question element extraction strategy is a summarization strategy, calling the BERT model to perform semantic summarization on each of the second question elements to obtain the target question element;

[0024] If the problem element extraction strategy is a precise strategy, each of the second problem elements is used as the target problem element.

[0025] Optionally, the matching and obtaining target after-sales knowledge data from the after-sales knowledge data set based on the target question elements includes:

[0026] The BERT model is used to perform semantic approximation analysis on the target question elements and each after-sales knowledge data in the after-sales knowledge data set to obtain corresponding approximation values, and the after-sales knowledge data with an approximation value higher than a threshold is determined as the target after-sales knowledge data.

[0027] The present invention also discloses a big data-based intelligent after-sales service provision system, which is applied to an intelligent customer service robot. The system includes a construction module, a first processing module, a second processing module, and a response module;

[0028] The construction module is used to obtain after-sales related big data related to the target brand, and construct a after-sales knowledge dataset based on the after-sales related big data;

[0029] The first processing module is used to receive first consulting information input by a target customer and determine a question element extraction strategy based on the first consulting information;

[0030] The second processing module is used to receive second consulting information input by the target customer, extract a plurality of second question elements from the second consulting information, construct a target question element based on each of the second question elements and the question element extraction strategy, and match the target after-sales knowledge data from the after-sales knowledge data set based on the target question element;

[0031] The response module is used to output the target after-sales knowledge to the target customer.

[0032] The present invention also discloses an electronic device, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement any of the above methods.

[0033] The present invention also discloses a computer storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement any of the above methods.

[0034] The present invention also discloses a computer program product, characterized in that: the computer program includes a number of computer codes, and when the computer codes are executed by a processor, any of the above methods is implemented.

[0035] The intelligent customer service robot in the solution of the present invention can determine a suitable problem element extraction strategy based on the characteristics of the target customer's general problem, and according to the problem element extraction strategy, appropriately integrate the various problem elements extracted and screened from the detailed after-sales demand content input by the target customer, thereby obtaining the target problem element, and then matching the target problem element to obtain the corresponding after-sales knowledge data. In this way, the intelligent customer service robot can generate a problem element set based on a suitable problem element extraction strategy, and filter the response content accordingly, without having to perform full element screening every time, which helps to increase the answer rate and reduce the customer's reply waiting time. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 It is a flowchart of a method for intelligently providing after-sales services based on big data disclosed in an embodiment of the present invention;

[0038] Figure 2 It is a schematic diagram of the interface of a technical discussion community / website targeted by the web crawler disclosed in the embodiment of the present invention;

[0039] Figure 3 It is a structural diagram of a system for intelligently providing after-sales services based on big data disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The following is a description of the implementation of the present application by specific specific embodiments. People familiar with the technology can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0041] In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0042] like Figure 1 As shown, the embodiment of the present invention discloses a method for intelligently providing after-sales service based on big data, which is applied to an intelligent customer service robot and includes the following steps:

[0043] Acquire after-sales related big data related to the target brand, and construct a after-sales knowledge dataset based on the after-sales related big data;

[0044] Receiving first consulting information input by a target customer, and determining a question element extraction strategy based on the first consulting information;

[0045] receiving second consulting information input by a target customer, extracting a plurality of second question elements from the second consulting information, constructing a target question element based on each of the second question elements and the question element extraction strategy, and matching the target after-sales knowledge data from the after-sales knowledge data set based on the target question element;

[0046] The target after-sales knowledge data is output to the target customer.

[0047] In the present invention, the target customer first simply inputs the first consulting information in the initial interface, and the intelligent customer service robot determines the problem element extraction strategy based on the first consulting information. Then, the target customer first inputs more detailed second consulting information in the formal service interface, and the intelligent customer service robot extracts several second problem elements from the second consulting information, and constructs a problem element set based on each second problem element and the aforementioned problem element extraction strategy, and performs an overall matching analysis from the after-sales knowledge data set based on the problem element set, so as to obtain the target after-sales knowledge. Finally, the target after-sales knowledge is output to the target customer. Among them, the after-sales knowledge data set of the present invention is constructed by the intelligent customer service robot based on the after-sales related big data related to the target brand, which can also reduce the cost of manually constructing the after-sales knowledge data set.

[0048] In the solution of the present invention, the intelligent customer service robot can determine the appropriate question element extraction strategy based on the characteristics of the target customer's general question, and according to the question element extraction strategy, appropriately integrate the various question elements extracted and screened from the detailed after-sales demand content input by the target customer, thereby obtaining the target question element, and then matching the target question element to obtain the corresponding after-sales knowledge data. With this arrangement, the intelligent customer service robot can generate a question element set based on an appropriate question element extraction strategy, and filter the response content accordingly, without having to perform full element screening every time, which helps to increase the answer rate and reduce the customer's response waiting time.

[0049] Optionally, the acquiring of after-sales related big data related to the target brand and constructing a after-sales knowledge dataset based on the after-sales related big data includes:

[0050] Determine product sequence information of the target brand, wherein the product sequence information includes a plurality of products arranged in a multi-level tree structure;

[0051] Using web crawler technology to crawl from the Internet to obtain after-sales related data corresponding to each of the products, each of the after-sales related data constitutes after-sales related big data;

[0052] Corresponding after-sales problems and corresponding solutions are extracted from each of the after-sales related data, and the after-sales problems are associated with the corresponding solutions to form an after-sales knowledge data. All the after-sales knowledge data are deduplicated, classified and integrated to obtain the after-sales knowledge data set.

[0053] In this embodiment, the construction of the existing after-sales knowledge data set is mainly based on manual construction. Some existing technologies use automatic construction technology, but this automatic construction technology is mainly based on the manually determined after-sales service manual. However, the after-sales knowledge data included in the after-sales service manual is often not complete, which will affect the accuracy of the intelligent customer service robot's answer to questions. In this regard, the present invention adopts network crawling technology to realize the automatic and rapid construction of the after-sales knowledge data set. The specific construction method is as follows:

[0054] First, determine the product sequence information of the target brand. The product sequence information contains multiple products under the brand. These products are arranged in a multi-level tree structure, such as range hoods - XX1 model, XX2 model; dishwashers - XX11 model, XX12 model, etc. Then, based on each product, use a web crawler program to crawl the Internet to obtain the corresponding after-sales related data. The crawler mainly targets the technical discussion communities / websites related to the product on the Internet (such as Figure 2As shown in the figure), short video platforms, etc. These network nodes contain a large number of fault problems and corresponding solutions. The crawled after-sales data constitutes after-sales related big data. Then, semantic analysis is performed on each after-sales related data to extract multiple groups of after-sales problems and solutions. Each group of after-sales problems and solutions constitutes a piece of after-sales knowledge data. After deduplication, classification and integration of the after-sales knowledge data, a classified after-sales knowledge data set can be integrated.

[0055] Among them, a web crawler is an automated program used to capture and process information from the Internet, mainly including a crawling engine, a scheduler, a URL manager, a downloader, and a parser. Its basic principle is to obtain web page content through network requests, and then parse and extract valuable data. Types of web crawlers include general web crawlers, focused web crawlers, incremental web crawlers, and deep web crawlers. General web crawlers crawl web pages extensively, focused web crawlers crawl specific topics, incremental web crawlers update crawled web pages, and deep web crawlers crawl dynamic web pages and database contents. The present invention can use the above-mentioned multiple types of crawlers to crawl after-sales related data at the same time, and then aggregate the crawled after-sales related data.

[0056] Optionally, the crawling of after-sales related data corresponding to each of the products from the Internet using a web crawler technology includes:

[0057] Accessing an online shopping website, and extracting from the online shopping website a first set of product detail pages of a target brand and a second set of product detail pages of several competing brands of the target brand;

[0058] Performing semantic analysis on the first set of product detail pages to extract a first number of promotional elements; and performing semantic analysis on the second set of product detail pages to extract a second number of promotional elements, calculating an absolute value of a difference between the first number and the second number, and determining a third number based on the absolute value of the difference;

[0059] Based on the third quantity, after-sales related data corresponding to each of the products is crawled from the Internet using web crawler technology.

[0060] In this embodiment, the amount of crawled after-sales related data should be appropriate. Too little data will make it difficult for the intelligent customer service robot to effectively answer some unpopular questions of customers; too much data will lead to a significant increase in computing power during subsequent integration and deduplication. In this regard, the present invention first analyzes the difference in the number of promotional elements between the target brand and the competing brand, and determines the appropriate crawling quantity based on this.

[0061] Specifically, the intelligent customer service robot visits some online shopping websites that mainly sell this type of product, and extracts the first set of product detail pages of the target brand and the second set of product detail pages of several competing brands of the target brand. By performing semantic analysis on the above product detail page set, the first number of promotional elements corresponding to the target brand and the second number of promotional elements corresponding to the competing brands can be extracted respectively, where the second number refers to the average number of promotional elements of all competing brands. Then, the absolute value of the difference between the first number and the second number obtained above is calculated, and the amount of after-sales related data that the web crawler should crawl for the specific product of the target brand, i.e., the third number, is determined based on the absolute value of the difference.

[0062] Among them, the number of the above-mentioned promotional elements actually represents the main selling point of a specific product of a specific brand, and the exclusive technology of the brand is often among them. Therefore, the present invention characterizes the probability that the target brand has a unique technology based on the absolute value of the difference between the number of selling points of the target brand and the average number of selling points of each competing brand. When the probability that the target brand has a unique technology is greater (that is, the greater the absolute value of the above-mentioned difference), it is set to crawl a small amount of after-sales related data of other competing brands after crawling the after-sales related data of the product, because the after-sales related data of other competing brands will not contain the after-sales data associated with the unique technology of the target brand, and the after-sales related data of the two are integrated. On the contrary, when the probability that the target brand has a unique technology is smaller (that is, the smaller the absolute value of the above-mentioned difference is), it is set to crawl more after-sales related data of other competing brands after crawling the after-sales related data of the product, which is conducive to extracting after-sales knowledge data corresponding to more similar after-sales problems, and integrating the after-sales related data of the two.

[0063] Therefore, through the above processing method, the present invention can achieve an appropriate amount of crawling of after-sales related data of each product of the target brand, thereby ensuring that the after-sales related data set finally constructed is of an appropriate scale, and an after-sales related data set of appropriate scale is conducive to the intelligent customer service robot to achieve a balance between answer accuracy and answer efficiency.

[0064] Optionally, the receiving first consultation information input by the target customer and determining a question element extraction strategy based on the first consultation information includes:

[0065] Displaying a plurality of question type buttons to the target customer, the buttons including a product category button and a plurality of product model buttons belonging to the product category button, receiving button selection information from the target customer, and determining the button selection information as the first consultation information of the target customer;

[0066] If the first consulting information corresponds to the product category button, the question element extraction strategy is set to a general strategy; if the first consulting information corresponds to the product model button, the question element extraction strategy is set to a precise strategy.

[0067] In this embodiment, after the target customer enters the after-sales service interface, the after-sales service interface first pops up buttons corresponding to different types of questions. These buttons are divided into two types: product category buttons and several product model buttons belonging to the product category buttons. Among them, the product category buttons refer to the product line types such as "gas stoves", "range hoods", and "dishwashers" under the target brand, and the product model buttons refer to the various product models under the above product line types, such as "gas stoves-11", "range hoods-21", "dishwashers-23", etc. Preferably, the product model buttons are the current hot-selling models. The target customer can make a choice according to his actual needs. If the target customer buys a conventional product, he will choose the product category button. If the target customer buys a hot-selling product, he will choose the corresponding product model button. Therefore, the above button selection information input by the target customer is determined as the first consulting information of the target customer, and the generalization strategy and the precision strategy are determined according to whether the first consulting information corresponds to the product category button or the product model button.

[0068] Therefore, by classifying the needs of target customers, different strategies for extracting question elements can be quickly determined, which helps to provide more appropriate personalized answers to target customers to improve their after-sales experience.

[0069] Optionally, extracting a plurality of second question elements from the second consulting information, and constructing a target question element based on each of the second question elements and the question element extraction strategy, includes:

[0070] Extracting a number of second question elements from the second consulting information;

[0071] If the question element extraction strategy is a summarization strategy, calling the BERT model to perform semantic summarization on each of the second question elements to obtain the target question element;

[0072] If the problem element extraction strategy is a precise strategy, each of the second problem elements is used as the target problem element.

[0073] In this embodiment, after the target customer selects each button in the interface as needed, the after-sales interaction stage is officially entered. The target customer enters the after-sales demand, i.e., the second consultation information, in the new interface, and the intelligent customer service robot performs element analysis on the second consultation information to extract several second question elements. If the aforementioned problem element extraction strategy determined is a summary strategy, the BERT model can be called to semantically summarize each second question element to obtain the summarized target question element; if the aforementioned problem element extraction strategy determined is a precise strategy, the extracted second question elements are directly used as the target question elements.

[0074] The following are some examples:

[0075] The second inquiry information input by the target customer is "I bought a XX brand water heater from a certain Dong on February 23, 24, model XX-32, and recently found that the hot water temperature was significantly lower than the set temperature value", and the second problem elements extracted from it are "water heater", "XX-32 model", and "low hot water temperature". If the problem element extraction strategy is a generalization strategy, the BERT model will re-understand the above second problem elements and perform a superordinate generalization, and the target problem element obtained is "poor heating performance of the water heater"; if the problem element extraction strategy is a precise strategy, "water heater", "XX-32 model", and "low hot water temperature" will be used as target problem elements.

[0076] Obviously, there are large differences in the target problem elements determined based on different problem element extraction strategies, and correspondingly, the target after-sales knowledge data obtained by subsequent matching will also be significantly different. The reason for the present invention to be set up in this way is that when the target customer selects the product category button, it means that the product he purchased is not a hot-selling model, and correspondingly, it is highly likely that it is not equipped with unique technology. At this time, the second problem elements are summarized, and multiple general solutions suitable for various types of "water heaters" are matched accordingly, which is more conducive to solving the target customer's fault problems, especially when the intelligent customer service robot outputs the target after-sales knowledge to each target customer to remotely guide them to perform self-service repairs, which may be faster and more effective. When the target customer selects the product model button, it means that the product he purchased is a hot-selling model, and correspondingly, the probability of it being equipped with unique technology is significantly greater. At this time, a full-factor search is performed based on each second problem element, and the specific target after-sales knowledge corresponding to the product can be searched and obtained, which is also conducive to guiding the target customer to perform remote self-service repairs. Of course, the target after-sales knowledge corresponding to the unique technology of hot-selling products may not be complete, and target customers can also try pressing the product category button at their own discretion. If the fault is not effectively resolved, they can press the product model button again.

[0077] It should be noted that a certain fault may be caused by conventional reasons or by the unique technology used. If it is caused by conventional reasons, there are likely to be multiple solutions obtained by matching (corresponding to the number of conventional reasons); if it is caused by the unique technology used, there are likely to be only one or a small number of solutions obtained by matching.

[0078] Optionally, the matching and obtaining target after-sales knowledge data from the after-sales knowledge data set based on the target question elements includes:

[0079] The BERT model is used to perform semantic approximation analysis on the target question elements and each after-sales knowledge data in the after-sales knowledge data set to obtain corresponding approximation values, and the after-sales knowledge data with an approximation value higher than a threshold is determined as the target after-sales knowledge data.

[0080] In this embodiment, BERT (Bidirectional Encoder Representations from Transformers) can learn the contextual representation of language through a deep bidirectional encoder, which is particularly suitable for deep analysis and understanding of natural language. In addition to using the BERT model to summarize the aforementioned second question elements into target question elements, the present invention also uses the BERT model to implement matching analysis, that is, using the BERT model to analyze the semantic similarity between the target question elements and each after-sales knowledge data, so as to match the best target after-sales knowledge data.

[0081] like Figure 3 As shown, the embodiment of the present invention also discloses a system for providing intelligent after-sales service based on big data, which is applied to an intelligent customer service robot. The system includes a building module, a first processing module, a second processing module, and a response module;

[0082] The construction module is used to obtain after-sales related big data related to the target brand, and construct a after-sales knowledge dataset based on the after-sales related big data;

[0083] The first processing module is used to receive first consulting information input by a target customer and determine a question element extraction strategy based on the first consulting information;

[0084] The second processing module is used to receive second consulting information input by the target customer, extract a plurality of second question elements from the second consulting information, construct a target question element based on each of the second question elements and the question element extraction strategy, and match the target after-sales knowledge data from the after-sales knowledge data set based on the target question element;

[0085] The response module is used to output the target after-sales knowledge to the target customer.

[0086] An embodiment of the present invention further discloses an electronic device, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the method described in the above embodiment.

[0087] An embodiment of the present invention further discloses a computer storage medium, wherein the computer storage medium stores a computer program, and the computer program is executed by a processor to implement the method described in the above embodiment.

[0088] The embodiment of the present invention further discloses a computer program product, characterized in that: the computer program includes a number of computer codes, and when the computer codes are executed by a processor, any of the above methods is implemented.

[0089] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0090] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0091] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for providing intelligent after-sales service based on big data, which is applied to an intelligent customer service robot, characterized in that: The method comprises the following steps: Acquire after-sales related big data related to the target brand, and construct a after-sales knowledge dataset based on the after-sales related big data; Receiving first consultation information input by a target customer, and determining a question element extraction strategy based on the first consultation information; receiving second consulting information input by a target customer, extracting a plurality of second question elements from the second consulting information, constructing a target question element based on each of the second question elements and the question element extraction strategy, and matching the target after-sales knowledge data from the after-sales knowledge data set based on the target question element; The target after-sales knowledge data is output to the target customer.

2. The method for providing intelligent after-sales service based on big data according to claim 1, characterized in that: Obtain after-sales related big data related to the target brand, and construct a after-sales knowledge dataset based on the after-sales related big data, including: Determine product sequence information of the target brand, wherein the product sequence information includes a plurality of products arranged in a multi-level tree structure; Using web crawler technology to crawl from the Internet to obtain after-sales related data corresponding to each of the products, each of the after-sales related data constitutes after-sales related big data; Corresponding after-sales problems and corresponding solutions are extracted from each of the after-sales related data, and the after-sales problems are associated with the corresponding solutions to form an after-sales knowledge data. All the after-sales knowledge data are deduplicated, classified and integrated to obtain the after-sales knowledge data set.

3. The method for providing intelligent after-sales service based on big data according to claim 2, characterized in that: Use web crawler technology to crawl from the Internet to obtain after-sales related data corresponding to each of the products, including: Accessing an online shopping website, and extracting from the online shopping website a first set of product detail pages of a target brand and a second set of product detail pages of several competing brands of the target brand; Performing semantic analysis on the first set of product detail pages to extract a first number of promotional elements; and performing semantic analysis on the second set of product detail pages to extract a second number of promotional elements, calculating an absolute value of a difference between the first number and the second number, and determining a third number based on the absolute value of the difference; Based on the third quantity, after-sales related data corresponding to each of the products is crawled from the Internet using web crawler technology.

4. The method for providing intelligent after-sales service based on big data according to claim 3 is characterized by: Receiving first consultation information input by a target customer, and determining a question element extraction strategy based on the first consultation information, including: Displaying a plurality of question type buttons to the target customer, the buttons including a product category button and a plurality of product model buttons belonging to the product category button, receiving button selection information from the target customer, and determining the button selection information as the first consultation information of the target customer; If the first consulting information corresponds to the product category button, the question element extraction strategy is set to a general strategy; if the first consulting information corresponds to the product model button, the question element extraction strategy is set to a precise strategy.

5. The method for providing intelligent after-sales service based on big data according to claim 4 is characterized by: Extracting a plurality of second question elements from the second consulting information, and constructing a target question element based on each of the second question elements and the question element extraction strategy, including: Extracting a number of second question elements from the second consulting information; If the question element extraction strategy is a summarization strategy, calling the BERT model to perform semantic summarization on each of the second question elements to obtain the target question element; If the problem element extraction strategy is a precise strategy, each of the second problem elements is used as the target problem element.

6. The method for providing intelligent after-sales service based on big data according to claim 1, characterized in that: The target after-sales knowledge data is obtained by matching from the after-sales knowledge data set based on the target problem elements, including: The BERT model is used to perform semantic approximation analysis on the target question elements and each after-sales knowledge data in the after-sales knowledge data set to obtain corresponding approximation values, and the after-sales knowledge data with an approximation value higher than a threshold is determined as the target after-sales knowledge data.

7. A smart after-sales service provision system based on big data, which is applied to an intelligent customer service robot. The system includes a construction module, a first processing module, a second processing module, and a response module; characterized in that: The construction module is used to obtain after-sales related big data related to the target brand, and construct a after-sales knowledge dataset based on the after-sales related big data; The first processing module is used to receive first consulting information input by a target customer and determine a question element extraction strategy based on the first consulting information; The second processing module is used to receive second consulting information input by the target customer, extract a plurality of second question elements from the second consulting information, construct a target question element based on each of the second question elements and the question element extraction strategy, and match the target after-sales knowledge data from the after-sales knowledge data set based on the target question element; The response module is used to output the target after-sales knowledge to the target customer.

8. An electronic device comprising: At least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the method according to any one of claims 1 to 6.

9. A computer storage medium storing a computer program, characterized in that: The computer program is executed by a processor to implement the method according to any one of claims 1 to 6.

10. A computer program product, characterized in that: The computer program includes a number of computer codes, and when the computer codes are executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images