A data knowledge base management system based on artificial intelligence

By introducing user pre-categorization and interest monitoring modules into the data knowledge base management system, the problems of user interest changes and data knowledge base type matching in the prior art are solved, and more accurate and personalized answer provision is achieved.

CN118410873BActive Publication Date: 2024-11-26JIANGSU AMP INFORMATION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410569852.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-09
Publication Date
2024-11-26
Estimated Expiration
2044-05-09

AI Technical Summary

Technical Problem

The prior art is difficult to cope with changes in user interests and correctly use the appropriate type of data knowledge base, thereby providing more accurate, personalized and answers that meet user needs.

Method used

Design a data knowledge base management system based on artificial intelligence, including data knowledge base module, user pre-classification module, intelligent question and answer module, monitoring module and evaluation module, pre-classify users through classification models, monitor changes in user interests, and according to Interest evaluation coefficient adjustment answer extraction strategy.

Benefits of technology

Through pre-categorization and interest monitoring, answers that are more in line with user needs are provided, ensuring that the system can be dynamically adjusted to adapt to changes in user interests and improve the accuracy and personalization of answers.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118410873B_ABST
    Figure CN118410873B_ABST
Patent Text Reader

Abstract

The present invention discloses a data knowledge base management system based on artificial intelligence, which specifically relates to the technical field of knowledge base management, including a data knowledge base module, a user pre-classification module, an intelligent question and answer module, a monitoring module and an evaluation module. According to different types of data, the data knowledge base is classified into multiple categories, a classification model is constructed using an intelligent algorithm, and users are classified into one or more knowledge base types according to their questions. According to the questions raised by the users, the system receives the questions and analyzes the questions to provide answers. According to the results of the user pre-classification, the answers are retrieved in the knowledge bases into which the users are classified, a monitoring interval is set, and within the monitoring interval, information on changes in user interests is collected according to the user's questions. The information on changes in interest is comprehensively analyzed to determine whether to change the answer extraction strategy. The present invention helps to select different data knowledge bases according to different users and provide answers that better meet the needs of users.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of knowledge base management, and more specifically, to a data knowledge base management system based on artificial intelligence. Background Art

[0002] In the prior art, after a user asks a question, the system first uses natural language processing technology to understand the user's intention and extract key information, then uses a machine learning model to classify the user into the corresponding knowledge base type, and then retrieves or generates the most relevant answer from the knowledge base based on text similarity matching or retrieval models. The system ensures that accurate and timely answers are provided based on the real-time updated knowledge base and model, but it is not easy to deal with retrieval problems caused by changes in user interests and correctly use the appropriate type of data knowledge base, thereby failing to provide more accurate, personalized answers that meet user needs.

[0003] In order to solve the above defects, a technical solution is now provided. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a data knowledge base management system based on artificial intelligence to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A data knowledge base management system based on artificial intelligence, including a data knowledge base module, a user pre-classification module, an intelligent question and answer module, a monitoring module and an evaluation module, and signal connections between the modules;

[0007] The data knowledge base module is used to classify the data knowledge base and classify the data knowledge base into multiple categories according to different types of data;

[0008] A user pre-classification module is used to build a classification model using an intelligent algorithm. The classification model classifies users into one or more knowledge base types based on their questions.

[0009] Intelligent question-answering module, which is used to receive questions raised by users based on artificial intelligence technology, analyze the questions and provide answers, and retrieve answers from the user's classified knowledge base based on the user's pre-classification results;

[0010] A monitoring module is used to set a monitoring interval, and collect user interest change information according to the user's questions within the monitoring interval, wherein the interest change information includes difference information obtained by analyzing the user's questions and periodic information of the user's own interest points;

[0011] The evaluation module is used to comprehensively analyze the interest change information, generate an interest evaluation coefficient, and determine whether to change the answer extraction strategy based on the comparison between the interest evaluation coefficient and the interest evaluation coefficient threshold.

[0012] In a preferred embodiment, collecting the user's interest change information according to the user's questions within the monitoring interval includes:

[0013] The interest change information is considered from two aspects, including analyzing the difference information of user questions and the periodic information of the user's interest points. The difference information is represented by the divergence change coefficient and the frequency deviation coefficient, and the periodic information is represented by the interest change coefficient.

[0014] In a preferred embodiment, the acquisition logic of the divergence variation coefficient is:

[0015] Obtain questions asked by users within the monitoring interval, determine the subject of each question asked by the user based on each question asked by the user, and determine the type of knowledge base to which the subject belongs;

[0016] By using the knowledge graph to determine the type of knowledge base to which the topic belongs, the topics and knowledge base types in the knowledge base are collected to build a knowledge graph. Each topic is a node in the graph, and each knowledge base type is also a node. The edge represents the association between the topic and the knowledge base type. Using the path search algorithm in the graph, the knowledge base type node related to the topic is found to determine the topic's affiliation. The topic's affiliation between different types of knowledge bases is compared, and the topic is classified as the knowledge base type with the highest affiliation.

[0017] According to the type of knowledge base to which the user is assigned, the type of knowledge base to which the user is assigned is recorded as a marked knowledge base, the number of topics belonging to the marked knowledge base in the monitoring interval is obtained, and the number of topics belonging to the marked knowledge base in the monitoring interval is marked as: DQ n , where n = 1, 2, 3, ..., N, N is a positive integer, and n is the type of the classified knowledge base;

[0018] The number of questions asked in the monitoring interval is obtained, and the number of questions asked in the monitoring interval is marked as: A, and the distribution of different types of knowledge bases in the marked knowledge base in the monitoring interval is calculated. The calculation formula is: Among them, Q n Mark the distribution of different types of knowledge bases in the knowledge base within the current monitoring interval;

[0019] Get the number of topics belonging to the labeled knowledge base during the pre-classification period, and mark the number of topics belonging to the labeled knowledge base during the pre-classification period as: ZQ n , calculate the distribution of different types of knowledge bases in the labeled knowledge base during pre-classification, and the calculation formula is: Among them, P nThe distribution of different types of knowledge bases in the labeled knowledge base during pre-classification;

[0020] The coefficient of variation of the divergence is calculated by the formula: Among them, SD bh is the coefficient of variation of the divergence.

[0021] In a preferred embodiment, the acquisition logic of the frequency deviation coefficient is:

[0022] Get the time length of the monitoring interval and mark the time length of the user monitoring interval as: SJ jc According to the number of topics belonging to the marked knowledge base in the monitoring interval, the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base in the monitoring interval is obtained, and the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base in the monitoring interval is marked as: JC n ,in,

[0023] Get the length of the pre-classification period and mark the length of the pre-classification period as: SJ fl , obtain the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base during the pre-classification period, and mark the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base during the pre-classification period as: FL n ,in,

[0024] The frequency deviation coefficient is calculated by the formula: Among them, PL pc is the frequency deviation coefficient.

[0025] In a preferred embodiment, the acquisition logic of the interest change coefficient is:

[0026] According to the type of knowledge base to which the user is assigned, topics included in different types of knowledge bases are obtained, topics that the user is interested in are obtained according to the historical records of user questions, and topics that the user is interested in are recorded as marked topics;

[0027] Collect the initial time when the user is interested in the marked topic, and obtain the time interval when the user is interested in the marked topic by collecting the time when the user is not interested in the marked topic, and mark the time interval when the user is interested in the marked topic as: JG k , where k = 1, 2, 3, ..., K, K is a positive integer, and k is the number of the marked topic;

[0028] Get the time average and standard deviation of the user's interest in the marked topic, and mark the time average and standard deviation of the user's interest in the marked topic as: JG avg and JGbzc ,in,

[0029] The interest change coefficient is calculated by the formula: XQ bh =JG bzc ×ln(JG bzc +1); among them, XQ bh is the coefficient of variation of interest.

[0030] In a preferred embodiment, the interest change information is comprehensively analyzed to generate an interest evaluation coefficient, including:

[0031] Comprehensively analyze the difference information of user questions and the periodic information of user interest points in the interest change information, establish a mathematical analysis model for the divergence change coefficient, frequency deviation coefficient and interest change coefficient, and generate an interest evaluation coefficient;

[0032] The divergence variation coefficient and interest variation coefficient are positively correlated with the interest evaluation coefficient, and the frequency deviation coefficient is negatively correlated with the interest evaluation coefficient.

[0033] In a preferred embodiment, determining whether to change the answer extraction strategy includes:

[0034] Set an interest evaluation coefficient threshold, and compare the interest evaluation coefficient in the monitoring interval with the interest evaluation coefficient threshold. If the interest evaluation coefficient is greater than the interest evaluation coefficient threshold, then based on the user's questions in the monitoring interval, re-use the classification model to determine the type of knowledge base to which the user is assigned. The system answers the user's questions based on the data stored in the knowledge base to which the user is assigned. If the interest evaluation coefficient is less than the interest evaluation coefficient threshold, continue to use the data in the marked knowledge base to answer the user's questions.

[0035] Technical effects and advantages of the present invention:

[0036] The present invention pre-classifies users by using a classification model to determine the areas in which users are initially interested. The system searches relevant data knowledge bases based on the areas in which users are initially interested, provides answers that users may need, and monitors interest changes in user questions based on multiple questions asked by users, and promptly changes answer extraction strategies based on changes in user interests, which helps to select different data knowledge bases based on different users and provide answers that better meet user needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0038] Figure 1The present invention is a structural schematic diagram of a data knowledge base management system based on artificial intelligence. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] Example 1

[0041] like Figure 1 A structural schematic diagram of a data knowledge base management system based on artificial intelligence of the present invention is given, which includes a data knowledge base module, a user pre-classification module, an intelligent question and answer module, a monitoring module and an evaluation module, and the modules are signal-connected;

[0042] The data knowledge base module is used to classify the data knowledge base and classify the data knowledge base into multiple categories according to different types of data;

[0043] A user pre-classification module is used to build a classification model using an intelligent algorithm. The classification model classifies users into one or more knowledge base types based on their questions.

[0044] Intelligent question-answering module, which is used to receive questions raised by users based on artificial intelligence technology, analyze the questions and provide answers, and retrieve answers from the user's classified knowledge base based on the user's pre-classification results;

[0045] A monitoring module is used to set a monitoring interval, and collect user interest change information according to the user's questions within the monitoring interval, wherein the interest change information includes difference information obtained by analyzing the user's questions and periodic information of the user's own interest points;

[0046] The evaluation module is used to comprehensively analyze the interest change information, generate an interest evaluation coefficient, and determine whether to change the answer extraction strategy based on the comparison between the interest evaluation coefficient and the interest evaluation coefficient threshold.

[0047] Classify the data knowledge base and classify it into multiple categories according to different types of data. Based on the classification of the data knowledge base, collect the user's question history data, including the content of the question, timestamp, etc., record the association between each question and the knowledge base type, determine the subject scope of the user's question, extract features from the user's question history data, train the extracted features into a classification model for classifying user types. Each user will be divided into one or more possible knowledge base types by the classification model. According to the knowledge base type to which the user is assigned, the system extracts the answers the user needs from the assigned knowledge base type.

[0048] It should be noted that the data knowledge base is classified according to different fields, and each knowledge base focuses on a specific field. When a user asks a question, the system selects the corresponding data knowledge base for retrieval based on the user's initial field of interest.

[0049] Among them, feature extraction is performed based on the user's question content, and the extracted features are used to train the model, including the following:

[0050] Extract keywords from the user's question content. Natural language processing technology can be used to extract keywords from the question content.

[0051] Extract topics from users’ questions and classify them into different topics using topic models or text classification algorithms;

[0052] Each feature is extracted to obtain a vector, the dimension of the total feature vector is obtained, and the total feature vector is used as the input of the classification model.

[0053] Users are classified according to the knowledge base type, and the classification results are expressed in the form of probability. Feature extraction is performed on the user's question history data, and the extracted features are trained into a classification model. The classification model is a probability-based multi-category classification task that outputs the probability distribution of each type by using a convolutional neural network.

[0054] The classification results of the classification model are displayed to the user through the system interface or other forms, so that the user can understand the type of knowledge base recommended by the system based on his or her question history, and encourage the user to provide feedback on the pre-classification results to continuously help optimize the classification model.

[0055] Based on the questions raised by the users, the system receives the questions, analyzes them and provides answers. Based on the results of the user's pre-classification, the system searches for answers in the classified knowledge base, provides appropriate answers based on the classified knowledge base, and prioritizes searching the knowledge base that matches the user's classification results. The system extracts the corresponding answers from the matching knowledge base and provides them to the users. These answers can be in the form of text, pictures, links, etc., depending on the nature of the user's question.

[0056] The answers are gradually updated based on the user's feedback on the provided answers. If the user is satisfied with the provided answer, the system keeps the way of providing the answer unchanged and uses such answers as a reference for providing answers in priority. If the user is not satisfied with the provided answer, the content of other types of knowledge bases is used to answer, and the previous answer is deleted to avoid providing it again and to prevent misleading the system to identify it as the correct answer.

[0057] Among them, answering users' questions is based on artificial intelligence technology. Based on the content of user questions, information retrieval technology is used to retrieve corresponding answers from the knowledge base, including building an inverted index to quickly locate documents or knowledge base entries containing user keywords, representing user questions and texts in the knowledge base as vectors, and calculating the similarity between them to find the most relevant answers. Based on semantic similarity measurement technology, such as word vector models, semantically related answers are found, etc., and natural language generation technology is used to automatically generate grammatically correct answers based on the retrieved content.

[0058] Monitor changes in user interest in their questions to avoid the system repeatedly searching for answers in the categorized knowledge base every time. When the user's interest changes, the user will needlessly repeat the search, which will reduce the efficiency of the system and increase unnecessary calculations and searches of the system.

[0059] A monitoring interval is set, which includes multiple questions asked by users. Through the questions asked by users and the answers given by the system within the monitoring interval, changes in the user's interest level are evaluated, and the monitoring interval can be dynamically adjusted according to the evaluation results. For example, when the user's interest changes slightly, the monitoring interval is kept unchanged and a larger range is maintained. When the user's interest changes greatly, the monitoring interval is shortened in time to monitor the user's interest changes more frequently. When the user's interest changes exceed a certain size, the user is reclassified based on the user's question history, and the knowledge base in which the user is classified is re-determined.

[0060] It should be noted that the monitoring interval is not a length of time, but represents the number of questions asked by users. When a user reaches a certain number of questions, the system evaluates the questions within the monitoring interval and determines changes in user interests. The monitoring interval is set by professional staff.

[0061] The user's interest change information is collected within the monitoring interval. The interest change information is considered from two aspects, including analyzing the difference information of the user's questions and the periodic information of the user's interest points. The difference information is represented by the divergence change coefficient and the frequency deviation coefficient. The divergence change coefficient represents the degree of change of the user's interest under the same number of questions, and the frequency deviation coefficient represents the degree of change of the user's interest per unit time. The periodic information is represented by the interest change coefficient, and the interest change coefficient is used to measure the persistence of the user's interest.

[0062] The acquisition logic of the divergence variation coefficient is: obtaining questions asked by users within the monitoring interval, determining the subject of each question asked by the user according to each question asked by the user, and determining the type of knowledge base to which the subject belongs;

[0063] Among them, for each question topic, the knowledge graph is used to determine the type of knowledge base to which the topic belongs, the topics and knowledge base types in the knowledge base are collected, and the knowledge graph is constructed. Each topic is a node in the graph, and each knowledge base type is also a node. The edge represents the association between the topic and the knowledge base type. The path search algorithm in the graph (such as the shortest path or minimum spanning tree algorithm) is used to find the knowledge base type node related to the topic, determine the topic's affiliation, compare the topic's affiliation between different types of knowledge bases, and classify the topic into the knowledge base type with the highest affiliation;

[0064] According to the knowledge base type to which the user is assigned, the assigned knowledge base type is recorded as the marked knowledge base, the number of topics belonging to the marked knowledge base in the monitoring interval is obtained, and the number of topics belonging to the marked knowledge base in the monitoring interval is marked as: DQ n , where n = 1, 2, 3, ..., N, N is a positive integer, and n is the type of the classified knowledge base;

[0065] It should be noted that the knowledge base belonging to the marked knowledge base is the type of knowledge base into which the user is classified. The topics asked by the user in the monitoring interval are divided into different knowledge base types according to the knowledge graph. Therefore, the number of topics belonging to the marked knowledge base can be obtained.

[0066] The number of questions asked in the monitoring interval is obtained, and the number of questions asked in the monitoring interval is marked as: A, and the distribution of different types of knowledge bases in the marked knowledge base in the monitoring interval is calculated. The calculation formula is: Among them, Q n Mark the distribution of different types of knowledge bases in the knowledge base within the current monitoring interval;

[0067] Get the number of topics belonging to the labeled knowledge base during the pre-classification period, and mark the number of topics belonging to the labeled knowledge base during the pre-classification period as: ZQ n, calculate the distribution of different types of knowledge bases in the labeled knowledge base during pre-classification, and the calculation formula is: Among them, P n The distribution of different types of knowledge bases in the labeled knowledge base during pre-classification;

[0068] It should be noted that the pre-classification period refers to the distribution of the topics of the user's questions in different types of knowledge bases when the initial user asks questions.

[0069] Calculate the coefficient of variation of the divergence using the formula: Among them, SD bh is the coefficient of variation of the divergence;

[0070] It can be seen from the formula that the larger the divergence variation coefficient is, the more different the topic types of users' questions in the monitoring interval are from those in the pre-classification period, indicating that users' interests may change significantly. Therefore, it may not be appropriate to answer users' questions using the knowledge base to which they are assigned.

[0071] The acquisition logic of the frequency deviation coefficient is: obtain the time length of the monitoring interval, and mark the time length of the user monitoring interval as: SJ jc According to the number of topics belonging to the marked knowledge base in the monitoring interval, the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base in the monitoring interval is obtained, and the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base in the monitoring interval is marked as: JC n ,in,

[0072] It should be noted that the monitoring interval is determined based on the number of questions asked by the user. The more questions a user asks within a certain period of time, the higher the frequency of monitoring changes in user interests.

[0073] Get the length of the pre-classification period and mark the length of the pre-classification period as: SJ fl , obtain the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base during the pre-classification period, and mark the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base during the pre-classification period as: FL n ,in,

[0074] Calculate the frequency deviation coefficient using the formula: Among them, PL pc is the frequency deviation coefficient.

[0075] It can be seen from the formula that the larger the frequency deviation coefficient, the more likely it is that the number of questions asked by users in the monitoring interval that are retrieved and answered using the marked knowledge base will increase, indicating that the user's interest concentration may deepen; the smaller the frequency deviation coefficient, the less likely it is that the number of questions asked by users in the monitoring interval that are retrieved and answered using the marked knowledge base will decrease, indicating that the user's interest has changed. Therefore, it may not be appropriate to use the knowledge base assigned to the user to answer the user's questions.

[0076] The logic for obtaining the interest change coefficient is as follows: according to the type of knowledge base to which the user is assigned, the topics included in the different types of knowledge bases are obtained, according to the historical records of the user's questions, the topics that the user is interested in are obtained, and the topics that the user is interested in are recorded as marked topics;

[0077] Collect the initial time when the user is interested in the marked topic, and use the user's first question about the marked topic as the initial time. Collect the time when the user is not interested in the marked topic, and obtain it through the user's feedback on the topic. Get the time interval when the user is interested in the marked topic, and mark the time interval when the user is interested in the marked topic as: JG k , where k = 1, 2, 3, ..., K, K is a positive integer, and k is the number of the marked topic;

[0078] Get the time average and standard deviation of the user's interest in the marked topic, and mark the time average and standard deviation of the user's interest in the marked topic as: JG avg and JG bzc ,in,

[0079] Calculate the interest change coefficient, the calculation formula is: XQ bh =JG bzc ×ln(JG bzc +1); among them, XQ bh is the coefficient of variation of interest.

[0080] It can be seen from the formula that the larger the interest change coefficient is, the more frequent or unstable the user's interest changes are, and the switching between topics of interest is more frequent. Therefore, it may not be appropriate to use the knowledge base to which the user is assigned to answer the user's questions.

[0081] The difference information of user questions and the periodic information of user interest points in the interest change information are comprehensively analyzed, and a mathematical analysis model is established for the divergence change coefficient, frequency deviation coefficient and interest change coefficient to generate the interest evaluation coefficient. The calculation formula of the interest evaluation coefficient is: Among them, pg xqis the interest assessment coefficient, α1, α2, α3 are the proportional coefficients of the divergence variation coefficient, the frequency deviation coefficient, and the interest variation coefficient, and α1, α2, α3 are greater than 0.

[0082] It can be seen from the formula that the smaller the frequency deviation coefficient, the larger the divergence change coefficient and the interest change coefficient, the larger the interest evaluation coefficient, indicating that the user may have a greater change in interest, and it may be less appropriate to use the knowledge base to which the user is assigned to answer the user's questions. Conversely, the larger the frequency deviation coefficient, the smaller the divergence change coefficient and the interest change coefficient, the smaller the interest evaluation coefficient, indicating that the user may have a smaller change in interest, and it may be more appropriate to use the knowledge base to which the user is assigned to answer the user's questions.

[0083] Set an interest evaluation coefficient threshold, and compare the interest evaluation coefficient in the monitoring interval with the interest evaluation coefficient threshold. If the interest evaluation coefficient is greater than the interest evaluation coefficient threshold, then based on the user's questions in the monitoring interval, re-use the classification model to determine the type of knowledge base to which the user is assigned. The system answers the user's questions based on the data stored in the knowledge base to which the user is assigned. If the interest evaluation coefficient is less than the interest evaluation coefficient threshold, continue to use the data in the marked knowledge base to answer the user's questions, that is, continue to use the previous answer extraction strategy.

[0084] This embodiment uses a classification model to pre-classify users and determine the areas of initial interest to the users. The system searches the relevant data knowledge base based on the areas of initial interest to provide answers that the users may need. It also monitors the interest changes of the users based on their multiple questions and promptly changes the answer extraction strategy based on the changes in the users' interests. This helps to select different data knowledge bases for different users and provide answers that better meet the users' needs.

[0085] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0086] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0087] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. 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.

[0088] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0089] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0090] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, 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 integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0091] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0092] The above is only a specific implementation 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 data knowledge base management system based on artificial intelligence, characterized in that: It includes a data knowledge base module, a user pre-classification module, an intelligent question-answering module, a monitoring module, and an evaluation module, and the modules are connected by signals; The data knowledge base module is used to classify the data knowledge base and classify the data knowledge base into multiple categories according to different types of data; A user pre-classification module is used to build a classification model using an intelligent algorithm. The classification model classifies users into one or more knowledge base types based on their questions. Intelligent question-answering module, which is used to receive questions raised by users based on artificial intelligence technology, analyze the questions and provide answers, and retrieve answers from the user's classified knowledge base based on the user's pre-classification results; A monitoring module is used to set a monitoring interval, and collect user interest change information according to the user's questions within the monitoring interval, wherein the interest change information includes difference information obtained by analyzing the user's questions and periodic information of the user's own interest points; An evaluation module is used to comprehensively analyze the interest change information, generate an interest evaluation coefficient, and determine whether to change the answer extraction strategy based on the comparison between the interest evaluation coefficient and the interest evaluation coefficient threshold; The user's interest change information is collected based on the user's questions during the monitoring period, including: The interest change information is considered from two aspects, including analyzing the difference information of the user's questions and the periodic information of the user's interest points. The difference information is represented by the divergence change coefficient and the frequency deviation coefficient, and the periodic information is represented by the interest change coefficient. The acquisition logic of the divergence variation coefficient is: Obtain questions asked by users within the monitoring interval, determine the subject of each question asked by the user based on each question asked by the user, and determine the type of knowledge base to which the subject belongs; By using the knowledge graph to determine the type of knowledge base to which the topic belongs, the topics and knowledge base types in the knowledge base are collected to build a knowledge graph. Each topic is a node in the graph, and each knowledge base type is also a node. The edge represents the association between the topic and the knowledge base type. Using the path search algorithm in the graph, the knowledge base type node related to the topic is found to determine the topic's affiliation. The topic's affiliation between different types of knowledge bases is compared, and the topic is classified as the knowledge base type with the highest affiliation. According to the knowledge base type to which the user is assigned, the assigned knowledge base type is recorded as the marked knowledge base, the number of topics belonging to the marked knowledge base in the monitoring interval is obtained, and the number of topics belonging to the marked knowledge base in the monitoring interval is marked as: DQ n , where n = 1, 2, 3, ..., N, N is a positive integer, and n is the type of the classified knowledge base; The number of questions asked in the monitoring interval is obtained, and the number of questions asked in the monitoring interval is marked as: A, and the distribution of different types of knowledge bases in the marked knowledge base in the monitoring interval is calculated. The calculation formula is: Among them, Q n Mark the distribution of different types of knowledge bases in the knowledge base within the current monitoring interval; Get the number of topics belonging to the labeled knowledge base during the pre-classification period, and mark the number of topics belonging to the labeled knowledge base during the pre-classification period as: ZQ n , calculate the distribution of different types of knowledge bases in the labeled knowledge base during pre-classification, and the calculation formula is: Among them, P n The distribution of different types of knowledge bases in the labeled knowledge base during pre-classification; The coefficient of variation of the divergence is calculated by the formula: Among them, SD bh is the coefficient of variation of the divergence; The acquisition logic of the frequency deviation coefficient is: Get the time length of the monitoring interval and mark the time length of the user monitoring interval as: SJ jc According to the number of topics belonging to the marked knowledge base in the monitoring interval, the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base in the monitoring interval is obtained, and the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base in the monitoring interval is marked as: JC n ,in, Get the length of the pre-classification period and mark the length of the pre-classification period as: SJ fl , obtain the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base during the pre-classification period, and mark the frequency of using different types of knowledge bases to retrieve answers belonging to the marked knowledge base during the pre-classification period as: FL n ,in, The frequency deviation coefficient is calculated by the formula: Among them, PL pc is the frequency deviation coefficient; The acquisition logic of the interest change coefficient is: According to the type of knowledge base to which the user is assigned, topics included in different types of knowledge bases are obtained, topics that the user is interested in are obtained according to the historical records of user questions, and topics that the user is interested in are recorded as marked topics; Collect the initial time when the user is interested in the marked topic, and obtain the time interval when the user is interested in the marked topic by collecting the time when the user is not interested in the marked topic, and mark the time interval when the user is interested in the marked topic as: JG k , where k = 1, 2, 3, ..., K, K is a positive integer, and k is the number of the marked topic; Get the time average and standard deviation of the user's interest in the marked topic, and mark the time average and standard deviation of the user's interest in the marked topic as: JG avg and JG bzc ,in, The interest change coefficient is calculated by the formula: XQ bh =JG bzc ×ln(JG bzc +1); among them, XQ bh is the coefficient of variation of interest; Comprehensively analyze interest change information to generate interest evaluation coefficients, including: Comprehensively analyze the difference information of user questions and the periodic information of user interest points in the interest change information, establish a mathematical analysis model for the divergence change coefficient, frequency deviation coefficient and interest change coefficient, and generate an interest evaluation coefficient; The divergence variation coefficient and interest variation coefficient are positively correlated with the interest evaluation coefficient, and the frequency deviation coefficient is negatively correlated with the interest evaluation coefficient.

2. According to the artificial intelligence-based data knowledge base management system of claim 1, it is characterized in that: Determine whether to change the answer extraction strategy, including: Set an interest evaluation coefficient threshold, and compare the interest evaluation coefficient in the monitoring interval with the interest evaluation coefficient threshold. If the interest evaluation coefficient is greater than the interest evaluation coefficient threshold, then based on the user's questions in the monitoring interval, re-use the classification model to determine the type of knowledge base to which the user is assigned. The system answers the user's questions based on the data stored in the knowledge base to which the user is assigned. If the interest evaluation coefficient is less than the interest evaluation coefficient threshold, continue to use the data in the marked knowledge base to answer the user's questions.