An enterprise knowledge Q&A analysis system and method based on an AI model

Abstract

The present invention discloses an enterprise knowledge Q&A analysis system and method based on an AI model, which relates to the technical field of AI models. Among them, the enterprise knowledge Q&A analysis method includes: analyzing the user's needs according to the user's question text, screening out the first recommended set of products, analyzing the functional matching between the historically purchased products and the products in the first recommended set according to the user's historical purchase record of products, screening out the second recommended set of products, analyzing the user's product usage habits according to the user's historical purchase record of products, analyzing the purchase suitability between the user and the products in the second recommended set according to the user's product usage habits, screening out the third recommended set of products, obtaining the corresponding product information in the third recommended set according to the user's historical purchase and browsing records, and recommending it to the user when answering questions. The present invention solves the technical problem in the prior art that the matching degree between product recommendation and user needs and user characteristics is not high, resulting in poor recommendation effects.

Description

Technical Field

[0001] The present invention relates to the technical field of AI models, and more particularly, to an enterprise knowledge Q&A analysis system and method based on an AI model. Background Art

[0002] An enterprise knowledge Q&A system is a tool based on artificial intelligence and natural language processing technologies, aiming to provide instant and accurate information answers for employees and customers;

[0003] However, in the related technologies, most are for making appropriate recommendations for products based on user questions, lacking a technical solution that jointly analyzes the user's needs, the functional matching of the products with the user's historical purchased products, and the user's usage habits and gives a recommended product that fits, resulting in a low degree of fit between the product recommendation and the user's needs and user characteristics, and being unable to solve the user's needs in a timely and accurate manner;

[0004] Regarding the technical problem in the prior art that the product recommendation has a low degree of matching with the user's questions and user characteristics, resulting in poor recommendation effects and being unable to solve the user's needs in a timely and accurate manner, no effective solution has been proposed yet;

[0005] Therefore, to solve the above problems, the present invention provides an enterprise knowledge Q&A analysis system and method based on an AI model. Summary of the Invention

[0006] The embodiments of the present invention provide an enterprise knowledge Q&A analysis system and method based on an AI model to at least solve the technical problem in the prior art that the product recommendation has a low degree of matching with the user's needs and user characteristics, resulting in poor recommendation effects and being unable to solve the user's needs in a timely and accurate manner.

[0007] According to one aspect of the embodiments of the present invention, the following technical solution is provided:

[0008] An enterprise knowledge Q&A analysis method based on an AI model includes the following specific steps:

[0009] S1. Analyze the user's needs based on the user's question text and screen out the first recommended set of products;

[0010] S2. Analyze the functional matching between the historically purchased products and the products in the first recommended set based on the user's historical purchased product records and screen out the second recommended set of products;

[0011] S3. Analyze the user's product usage habits based on the user's historical purchased product records;

[0012] S4. Analyze the purchase suitability between the user and the products in the second recommended set based on the user's product usage habits and screen out the third recommended set of products;

[0013] S5. Obtain the corresponding product information in the third recommendation set based on the user's historical purchase and browsing records, and recommend it to the user when answering questions.

[0014] Optionally, the S1 includes the following specific steps:

[0015] S101. Input the user's question text into the language processing module, and extract the user demand keywords in the question text. The user demand keywords include demand individual keywords and service subject keywords.

[0016] S102. Screen out the products containing the demand individual keywords in the enterprise product management module according to the demand individual keywords and output them in the form of a first recommendation set.

[0017] Optionally, the S2 includes the following specific steps:

[0018] S201. Obtain the user's historical purchased product records, analyze the keywords of the historical purchased products according to the historical purchased product records. The keywords of the historical purchased products include name keywords and function keywords. Screen out the historical purchased products containing the service subject keywords according to the name keywords and the service subject keywords, and at the same time obtain the function keywords of the products in the first recommendation set.

[0019] S202. Analyze the function matching degree between the screened historical purchased products and the products in the first recommendation set according to the function keywords of the screened historical purchased products and the function keywords of the products in the first recommendation set.

[0020] S203. Sort the function matching degrees in descending order, screen out the products corresponding to the set serial numbers and output them in the form of a second recommendation set.

[0021] Optionally, the S3 includes the following specific steps:

[0022] S301. Extract the purchase time and purchase amount from the screened historical purchased product records, obtain the purchase time difference between adjacent two historical purchased product records, then take the average value of the obtained several purchase time differences to get the average purchase time difference, and take the average value of the obtained several purchase amounts to get the average purchase amount.

[0023] S302. Compare the average purchase time difference with a preset first time difference threshold and a second time difference threshold to determine the user's usage habit. If the average purchase time difference is less than or equal to the first time difference threshold, it is determined that the user has a high frequency of replacing the commodity. If the average purchase time difference is greater than the first time difference threshold and less than the second time difference threshold, it is determined that the user has a medium frequency of replacing the commodity. If the average purchase time difference is greater than or equal to the second time difference threshold, it is determined that the user has a low frequency of replacing the commodity, where the first time difference threshold is less than the second time difference threshold.

[0024] S303. Obtain a user usage habit score value according to the user's usage habit.

[0025] Optionally, the S4 includes the following specific steps:

[0026] S401. Obtain the usage keywords and commodity amounts of the commodities in the second recommendation set, and analyze the purchase suitability between the user and the commodities in the second recommendation set according to the user usage habit score value, the average purchase amount, the usage keywords and the commodity amounts of the commodities in the second recommendation set.

[0027] S402. Sort the purchase suitability in descending order, screen out the corresponding commodities within the set serial number, and output them in the form of a third recommendation set.

[0028] Optionally, the S5 includes the following specific steps:

[0029] Obtain the user's historical purchase and browsing records to analyze whether the user has an analogy habit. If the user does not have an analogy habit, obtain the commodity information with the highest usage matching degree in the third recommendation set and recommend it to the user when answering questions. If the user has an analogy habit, obtain the corresponding commodity information within the set serial number in the third recommendation set and recommend it to the user when answering questions.

[0030] According to another aspect of the embodiments of the present invention, there is also provided an enterprise knowledge Q&A analysis system based on an AI model for implementing an enterprise knowledge Q&A analysis method based on an AI model, including: a requirement analysis unit for inputting a user question text into a language processing module, extracting user requirement keywords from the question text, and screening out the commodities containing the requirement individual keywords in the enterprise commodity management module and outputting them in the form of a first recommendation set.

[0031] A function matching analysis unit, configured to obtain the user's historical purchased product records, analyze the keyword of the historical purchased products according to the historical purchased product records, screen out the historical purchased products containing the service subject keyword according to the name keyword and the service subject keyword, and analyze the function matching degree between the screened historical purchased products and the products in the first recommended set according to the function keyword of the screened historical purchased products and the function keyword of the products in the first recommended set;

[0032] A usage habit analysis unit, configured to extract the purchase time and purchase amount from the screened historical purchased product records, compare the average purchase time difference with the preset first time difference threshold and second time difference threshold to judge the user's usage habit, and obtain the user usage habit score value according to the user's usage habit;

[0033] A purchase adaptation analysis unit, configured to obtain the usage keyword and product amount of the products in the second recommended set, and analyze the purchase adaptation degree between the user and the products in the second recommended set according to the user usage habit score value, the average purchase amount, the usage keyword and the product amount of the products in the second recommended set;

[0034] A product recommendation unit, configured to obtain the user's historical purchase and browsing records to analyze whether the user has an analogy habit. If the user does not have an analogy habit, obtain the product information with the highest usage matching degree in the third recommended set and recommend it to the user when answering questions. If the user has an analogy habit, obtain the corresponding product information within the set serial number in the third recommended set and recommend it to the user when answering questions.

[0035] According to another aspect of the embodiments of the present invention, there is also provided a computer-readable storage medium, which is used to store a computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute an enterprise knowledge Q&A analysis method based on an AI model as described in any one of the above.

[0036] According to another aspect of the embodiments of the present invention, there is also provided an electronic device, including one or more processors and a memory. The memory is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement an enterprise knowledge Q&A analysis method based on an AI model as described in any one of the above.

[0037] In the present invention, the user's needs are analyzed based on the user's question text, and the first recommended set of products is screened out. The functional matching between the historically purchased products and the products in the first recommended set is analyzed according to the user's historical purchase record of products, and the second recommended set of products is screened out. The user's product usage habits are analyzed according to the user's historical purchase record of products, and the purchase suitability between the user and the products in the second recommended set is analyzed according to the user's product usage habits, and the third recommended set of products is screened out. The corresponding product information in the third recommended set is obtained according to the user's historical purchase and browsing records and recommended to the user when answering the question. The present invention combines the user's needs, the functional matching with the user's historically purchased products, and the user's usage habits for joint analysis and gives recommended products that are suitable, which is beneficial to improving the suitability between the products and the user, timely and accurately solving the user's needs, and improving the enterprise service satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The drawings described herein are used to provide a further understanding of the present invention and form a part of this application. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0039] Figure 1 It is a schematic flow diagram of a method for enterprise knowledge question-answering analysis based on an AI model of the present invention;

[0040] Figure 2 It is a schematic flow diagram of process S1 in a method for enterprise knowledge question-answering analysis based on an AI model of the present invention;

[0041] Figure 3 It is a schematic flow diagram of process S2 in a method for enterprise knowledge question-answering analysis based on an AI model of the present invention;

[0042] Figure 4 It is a schematic flow diagram of process S3 in a method for enterprise knowledge question-answering analysis based on an AI model of the present invention;

[0043] Figure 5 It is a schematic flow diagram of process S4 in a method for enterprise knowledge question-answering analysis based on an AI model of the present invention;

[0044] Figure 6 It is a schematic overall framework diagram of a system for enterprise knowledge question-answering analysis based on an AI model of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] To enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0046] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0047] To facilitate the understanding of the present invention by those skilled in the art, the following explains some terms or nouns related to each embodiment of the present invention:

[0048] AI model: An AI model refers to a model obtained by training with a large amount of data based on deep learning algorithms, which can perform various tasks, such as natural language processing, image recognition, speech recognition, etc. By learning the features and patterns in the data, the AI model can perform complex tasks and has made remarkable progress in various fields.

[0049] AI models can be classified according to the model structure into:

[0050] Deep neural network (DNN), generative adversarial network (GAN), variational autoencoder (VAE);

[0051] AI models can be classified according to the task type into:

[0052] Natural language processing (NLP) models, computer vision (CV) models, multi-modal models;

[0053] The fields of application of AI models include:

[0054] Natural language processing: Used for text generation, question answering systems, machine translation, etc.;

[0055] Computer vision: Used for image recognition, video analysis, object detection, etc.;

[0056] Speech recognition and generation: used in voice assistants, speech-to-text, etc.;

[0057] Recommendation system: used in personalized recommendations, advertising, etc.;

[0058] Bioinformatics: used in gene sequence analysis, protein structure prediction, etc.

[0059] NLP: Natural Language Processing (NLP) is an important research direction in the field of artificial intelligence. It integrates knowledge from multiple disciplinary fields such as linguistics, computer science, machine learning, mathematics, and cognitive psychology. It is an interdisciplinary subject that combines computer science, artificial intelligence, and linguistics. It includes two main aspects: natural language understanding and natural language generation. The research content includes various levels such as characters, words, phrases, sentences, paragraphs, and texts. It is a bridge for communication between machine language and human language. Its purpose is to enable machines to understand, interpret, and generate human language, achieve effective communication between humans and machines, and enable computers to perform tasks such as language translation, sentiment analysis, and text summarization.

[0060] The present invention will be described in detail below with reference to each embodiment.

[0061] Embodiment 1

[0062] According to an embodiment of the present invention, an embodiment of an enterprise knowledge Q&A analysis method based on an AI model is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0063] Figure 1 is a flowchart of an optional enterprise knowledge Q&A analysis method based on an AI model according to an embodiment of the present invention, as Figure 1 shown, the method includes the following steps:

[0064] S1. Analyze the user's needs based on the user's question text and screen out the first recommended set of products;

[0065] As Figure 2 shown, in this embodiment, S1 includes the following specific steps:

[0066] S101. Input the user's question text into the language processing module, and extract the user's need keywords from the question text. The user's need keywords include need individual keywords and service subject keywords;

[0067] Optionally, the user's question is analyzed by an AI model, and the language processing module uses NLP algorithms for text preprocessing. The text preprocessing steps include: removing irrelevant characters (deleting special characters, punctuation marks, etc. in the text), word segmentation (dividing the text into words or phrases), and removing stop words (deleting high-frequency but less meaningful words in the text). Then, the TF-IDF (Term Frequency-Inverse Document Frequency) algorithm is used to calculate the TF-IDF values, and the words are sorted according to the scores output by the model. The words within the set sequence range are selected as keywords, or a deep learning model (such as TextRank, Word2Vec, BERT) is used for feature extraction to obtain the user demand keywords. The user demand keywords include demand individual keywords and service entity keywords. The demand individual is the commodity that the user needs this time according to the question in the user's question, and the service entity is the commodity that the obtained commodity serves. For example, if the user asks in the enterprise knowledge Q&A module whether there is a safety seat for infants during driving, the demand individual keyword is the safety seat, and the service individual keyword is the vehicle.

[0068] S102. Screen out the commodities containing the demand individual keywords in the enterprise commodity management module and output them in the form of a first recommended set.

[0069] S2. Analyze the functional matching degree between the historical purchased commodities and the commodities in the first recommended set according to the user's historical purchased commodity records, and screen out a second recommended set of commodities;

[0070] As Figure 3 shown, in this embodiment, S2 includes the following specific steps:

[0071] S201. Obtain the user's historical purchased commodity records, analyze the keywords of the historical purchased commodities according to the historical purchased commodity records. The keywords of the historical purchased commodities include name keywords and functional keywords. Screen out the historical purchased commodities containing the service entity keywords according to the name keywords and service entity keywords, and at the same time obtain the functional keywords of the commodities in the first recommended set;

[0072] S202. Analyze the functional matching degree between the screened historical purchased commodities and the commodities in the first recommended set according to the functional keywords of the screened historical purchased commodities and the functional keywords of the commodities in the first recommended set. Among them, the functional matching degree between the screened historical purchased commodities and the commodities in the first recommended set can be obtained through the functional matching degree calculation formula. The functional matching degree calculation formula can be: , where is the vector of the i-th functional keyword of the I-th screened historical purchased commodity, is the vector of the j-th functional keyword of the J-th product in the first recommended set. Among them, a neural network is trained using the continuous bag-of-words model (CBOW) or the skip-gram model to learn the vector representation of the keywords, so that semantically similar words are close in the vector space, and the vector representation of the keywords is output. is the dot product. is the modulus of the vector of the i-th functional keyword of the I-th filtered historical purchased product. is the modulus of the vector of the j-th functional keyword of the J-th product in the first recommended set. is the weight of the i-th functional keyword of the I-th filtered historical purchased product. is the weight of the j-th functional keyword of the J-th product in the first recommended set. The score output by the model when obtaining the functional keyword is imported into the pre-optimized fitting software, and the fitting software outputs the matching weight, that is and the values of;

[0073] Optionally, obtain the relevant text information of the filtered historical purchased products and the products in the first recommended set, including product titles, product details, product reviews, etc. After cleaning the relevant text information, use a deep learning model to extract features from the text to obtain the keywords representing the product functions. For example, the functional keywords of a safety seat are: "ISOFIX interface", "earthquake resistance", "7-level adjustment", "suitable for 0-4 years old"; the functional keywords of a vehicle are: "ISOFIX interface", "spacious rear seat space", "child safety lock".

[0074] S203. Sort the functional matching degrees in descending order, filter out the corresponding products within the set serial numbers, and output them in the form of a second recommended set.

[0075] S3. Analyze the user's product usage habits based on the user's historical purchased product records;

[0076] As Figure 4 shown, in this embodiment, S3 includes the following specific steps:

[0077] S301. Extract the purchase time and purchase amount from the filtered historical purchased product records, obtain the time difference between the purchase times of adjacent two historical purchased product records, then take the average of the obtained several time differences to get the average purchase time difference, and take the average of the obtained several purchase amounts to get the average purchase amount.

[0078] S302. Compare the average purchase time difference with a preset first time difference threshold and a second time difference threshold, and determine the user's usage habit. If the average purchase time difference is less than or equal to the first time difference threshold, it is determined that the user has a high frequency of replacing the product. If the average purchase time difference is greater than the first time difference threshold and less than the second time difference threshold, it is determined that the user has a medium frequency of replacing the product. If the average purchase time difference is greater than or equal to the second time difference threshold, it is determined that the user has a low frequency of replacing the product. Herein, the first time difference threshold is less than the second time difference threshold;

[0079] S303. Obtain a user usage habit score value according to the user's usage habit. Specifically, hire relevant experts to analyze the user's usage habit and give a usage habit score value. Input the usage habit score value given by the expert into the fitting software for optimization. Import the user's usage habit into the pre-optimized fitting software, and let the fitting software output a matching user usage habit score value.

[0080] S4. Analyze the purchase suitability between the user and the products in the second recommended set according to the user's product usage habit, and screen out a third recommended set of products;

[0081] As Figure 5 shown, in this embodiment, S4 includes the following specific steps:

[0082] S401. Obtain the usage keywords and product amounts of the products in the second recommended set, and analyze the purchase suitability between the user and the products in the second recommended set according to the user usage habit score value, the average purchase amount, the usage keywords and product amounts of the products in the second recommended set. Herein, the purchase suitability between the user and the products in the second recommended set can be obtained through a purchase suitability calculation formula. The purchase suitability calculation formula can be: , wherein is the amount of the p-th product in the second recommended set, is the average purchase amount. Analyze whether the p-th product in the second recommended set matches the user's consumption habit through . When the value is larger, it indicates that the product matches the user's consumption habit better. is the modulus of the k-th usage keyword vector of the product, is the weight of the k-th usage keyword pair of the product, is the user usage habit score value. Analyze whether each product matches the user's usage habit through . When the value is larger, it indicates that the product matches the user's usage habit better. Herein, import the usage keywords of the products in the second recommended set into the pre-optimized fitting software, and let the fitting software output the matching weight, that is, the value of ;

[0083] S402. Sort the purchase fitness in descending order, filter out the corresponding products within the set serial numbers, and output them in the form of a third recommended set.

[0084] S5. Obtain the corresponding product information in the third recommended set according to the user's historical purchase and browsing records, and recommend it to the user when answering questions.

[0085] In this embodiment, S5 includes the following specific steps:

[0086] Obtain the user's historical purchase and browsing records to analyze whether the user has an analogy habit. If the user does not have an analogy habit, obtain the product information with the highest usage matching degree in the third recommended set and recommend it to the user when answering questions. If the user has an analogy habit, obtain the corresponding product information within the set serial numbers in the third recommended set and recommend it to the user when answering questions.

[0087] Optionally, obtain the user's historical browsing records, including product categories, browsing times, click times, etc., extract the user's behavior characteristics, and analyze whether the user tends to frequently browse similar products when wanting to purchase a certain type of product. If the user frequently browses similar products, it is determined that the user has an analogy habit.

[0088] The following describes the present invention in conjunction with another optional embodiment.

[0089] Embodiment 2

[0090] The embodiment of the present invention provides an enterprise knowledge Q&A analysis system based on an AI model. Each implementation unit included in the enterprise knowledge Q&A analysis system corresponds to each implementation step in the first embodiment above.

[0091] Figure 6 It is a schematic diagram of an optional enterprise knowledge Q&A analysis system based on an AI model according to the embodiment of the present invention. As Figure 6 shown, the enterprise knowledge Q&A analysis system may include: a requirement analysis unit, configured to input the user's question text into the language processing module, extract the user's requirement keywords from the question text, and filter out the products containing the requirement individual keywords in the enterprise product management module according to the requirement individual keywords and output them in the form of a first recommended set;

[0092] A function matching analysis unit, configured to obtain the user's historical purchased product records, analyze the keywords of the historical purchased products according to the historical purchased product records, filter out the historical purchased products containing the service subject keywords according to the name keywords and service subject keywords, and analyze the function matching degree between the filtered historical purchased products and the products in the first recommended set according to the function keywords of the filtered historical purchased products and the products in the first recommended set;

[0093] A usage habit analysis unit, configured to extract the purchase time and purchase amount from the filtered historical purchased commodity records, compare the average purchase time difference with a preset first time difference threshold and a second time difference threshold to determine the user's usage habit, and obtain a user usage habit score value according to the user's usage habit;

[0094] A purchase adaptation analysis unit, configured to obtain the usage keywords and commodity amounts of the commodities in the second recommendation set, and analyze the purchase adaptability between the user and the commodities in the second recommendation set according to the user usage habit score value, the average purchase amount, the usage keywords and the commodity amounts of the commodities in the second recommendation set;

[0095] A commodity recommendation unit, configured to analyze whether the user has an analogy habit based on the user's historical purchase and browsing records. If the user does not have an analogy habit, obtain the commodity information with the highest usage matching degree in the third recommendation set and recommend it to the user when answering questions. If the user has an analogy habit, obtain the corresponding commodity information within the set serial number in the third recommendation set and recommend it to the user when answering questions.

[0096] Embodiment 3

[0097] According to another aspect of the embodiments of the present invention, there is also provided a computer-readable storage medium, which is used to store a computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute an AI model-based enterprise knowledge Q&A analysis method according to any one of the above.

[0098] Embodiment 4

[0099] According to another aspect of the embodiments of the present invention, there is also provided an electronic device, including one or more processors and a memory. The memory is used to store one or more programs. When one or more programs are executed by one or more processors, one or more processors are enabled to implement an AI model-based enterprise knowledge Q&A analysis method according to any one of the above.

[0100] This electronic device may vary greatly due to configuration or performance differences. It can include one or more processors (Central Processing Units, CPUs) and one or more memories. Among them, at least one computer program is stored in the memory, and the computer program is loaded and executed by the processor to implement an AI model-based enterprise knowledge Q&A analysis method provided by the above method embodiments. This electronic device can also include other components for implementing the functions of the device. For example, this electronic device can also have components such as wired or wireless network interfaces and input / output interfaces for data input and output. This embodiment will not be elaborated here.

[0101] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0102] In the above embodiments of the present invention, the descriptions of the respective embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0103] In several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the division of units can be a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of units or modules can be in an electrical or other form.

[0104] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0105] In addition, the functional units in the various embodiments of the present invention can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0106] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the methods in the various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, read-only memories (ROMs), random access memories (RAMs), mobile hard disks, magnetic disks or optical disks and other various media that can store program codes.

[0107] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. An enterprise knowledge question-answering analysis method based on an AI model, characterized in that: The specific steps include: S1. Analyze user needs based on the user's question text and select the first recommended set of products; S2. Analyze the functional matching between the historically purchased products and the products in the first recommended set according to the historical purchase records of the user, and select a second recommended set of products; S3. Analyze the user's commodity usage habits based on the user's historical commodity purchase records; S4. Analyze the purchasing compatibility of the user with the products in the second recommendation set according to the user's product usage habits, and select a third recommendation set of products; S5, obtaining corresponding product information in the third recommendation set according to the user's historical purchase and browsing records, and recommending it to the user when answering questions; S1 includes the following specific steps: S101, inputting a user question text into a language processing module, and extracting user demand keywords from the question text, wherein the user demand keywords include demand individual keywords and service subject keywords; S102, screening out commodities containing the individual demand keywords in the enterprise commodity management module according to the individual demand keywords and outputting them in the form of a first recommendation set; S2 includes the following specific steps: S201, obtaining a user's historical purchase records, analyzing keywords of historical purchases according to the historical purchase records, wherein the keywords of the historical purchases include name keywords and function keywords, filtering out historical purchases containing service subject keywords according to the name keywords and the service subject keywords, and obtaining function keywords of the products in the first recommendation set; S202, analyzing the functional matching degree between the filtered historically purchased products and the products in the first recommended set according to the functional keywords of the filtered historically purchased products and the functional keywords of the products in the first recommended set; S203, sorting the function matching degrees in descending order, filtering out the products corresponding to the set serial numbers and outputting them in the form of a second recommendation set; The S5 comprises the following specific steps: Obtain the user's historical purchase and browsing records to analyze whether the user has the habit of analogy. If the user does not have the habit of analogy, obtain the product information with the highest matching value in the third recommendation set and recommend it to the user when answering questions. If the user has the habit of analogy, obtain the product information corresponding to the set serial number in the third recommendation set and recommend it to the user when answering questions.

2. The enterprise knowledge question-answering analysis method based on an AI model as claimed in claim 1, characterized in that: The S3 comprises the following specific steps: S301, extracting the purchase time and purchase amount from the filtered historical purchase records, obtaining the purchase time difference between two adjacent historical purchase records, then averaging several obtained purchase time differences to obtain an average purchase time difference, and averaging several obtained purchase amounts to obtain an average purchase amount; S302, comparing the average purchase time difference with a preset first time difference threshold and a second time difference threshold and judging the user's usage habits, if the average purchase time difference is less than or equal to the first time difference threshold, judging that the user changes the goods frequently, if the average purchase time difference is greater than the first time difference threshold and the purchase time difference is less than the second time difference threshold, judging that the user changes the goods moderately, if the average purchase time difference is greater than or equal to the second time difference threshold, judging that the user changes the goods slowly; S303: Obtain a user usage habit score value based on the user's usage habits.

3. The enterprise knowledge question-answering analysis method based on an AI model as claimed in claim 2, characterized in that: The S4 comprises the following specific steps: S401, obtaining usage keywords and commodity amounts of commodities in the second recommendation set, and analyzing the purchase compatibility between the user and the commodities in the second recommendation set according to the user usage habit score, the average purchase amount, the usage keywords and commodity amounts of the commodities in the second recommendation set; S402: Sort the purchase suitability in descending order, filter out the commodities corresponding to the set serial numbers, and output them in the form of a third recommendation set.

4. An enterprise knowledge question and answer analysis system based on an AI model, used to implement the enterprise knowledge question and answer analysis method based on an AI model as claimed in any one of claims 1 to 3, characterized in that: include: The demand analysis unit is used to input the user's question text into the language processing module, extract the user's demand keywords from the question text, and select the products containing the demand individual keywords in the enterprise product management module according to the demand individual keywords and output them in the form of a first recommendation set; a function matching analysis unit, configured to obtain a user's historical purchase records, analyze keywords of historical purchases according to the historical purchase records, filter out historical purchases containing the service subject keywords according to the name keywords and the service subject keywords, and analyze the function matching degree between the filtered historical purchases and the commodities in the first recommendation set according to the function keywords of the filtered historical purchases and the function keywords of the commodities in the first recommendation set; A usage habit analysis unit is used to extract the purchase time and purchase amount from the filtered historical purchase records, compare the average purchase time difference with the preset first time difference threshold and the second time difference threshold, and judge the user's usage habits, and obtain the user's usage habit score value according to the user's usage habits; A purchase adaptation analysis unit, used to obtain the usage keywords and commodity amounts of the commodities in the second recommendation set, and analyze the purchase compatibility between the user and the commodities in the second recommendation set according to the user usage habit score, the average purchase amount, the usage keywords and the commodity amounts of the commodities in the second recommendation set; The product recommendation unit is used to obtain the user's historical purchase and browsing records to analyze whether the user has an analogy habit. If the user does not have an analogy habit, the product information with the highest matching probability in the third recommendation set is obtained and recommended to the user when answering questions. If the user has an analogy habit, the product information corresponding to the set serial number in the third recommendation set is obtained and recommended to the user when answering questions.

5. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store a computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute an enterprise knowledge question and answer analysis method based on an AI model as described in any one of claims 1 to 3.

6. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement an enterprise knowledge question and answer analysis method based on an AI model as described in any one of claims 1 to 3.

Images