An intelligent question-answering robot system based on online knowledge community

Abstract

The invention discloses an intelligent question-answering robot system based on an online knowledge community, including a user information module, a data interaction module, an intelligent question-answering system, a knowledge base management module and a learning and optimization module, and the invention relates to the field of artificial intelligence technology. The intelligent question-answering robot system based on an online knowledge community, by being provided with an intelligent question-answering system, adopts advanced natural language processing technology and artificial intelligence algorithms, realizes automatic parsing and understanding of user input, and completes processing of natural language questions expressed by users through a natural language processing module, completes text cleaning and corresponding conjunction vocabulary feature extraction to improve the accuracy of understanding natural language, and sorts and combines conjunction vocabulary features to improve the accuracy of reply results.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an intelligent question-answering robot system based on an online knowledge community. Background Art

[0002] The intelligent question-answering robot system based on online knowledge communities involves multiple aspects such as natural language processing technology, knowledge representation and reasoning technology, deep learning technology, and the platform advantages of online knowledge communities. The combination of these technologies enables the intelligent question-answering robot to accurately understand, quickly retrieve, and accurately answer user questions, providing users with a better service experience.

[0003] The reference patent name is: An intelligent robot automatic question and answer system (patent publication number: CN115952262A, patent publication date: 2023-04-11), which belongs to the field of human-computer intelligent interaction technology, including a database, which is used to store question and answer data; it also includes an interaction module, a question and answer module, an adjustment module and a server; the interaction module is used for users to input questions, and perform domain analysis on the questions input by users, obtain question retrieval data, and send the obtained question retrieval data to the question and answer module; the question and answer module analyzes the received question retrieval data, obtains answers to the questions, and sends the obtained answers to the questions to the interaction module; the interaction module sends the received answers to the questions to the user, and receives the user's feedback data on the answers to the questions, and sends the obtained feedback data to the adjustment module, the adjustment module analyzes the received feedback data, and adjusts the domain analysis of the user's questions in the interaction module and the question and answer in the question and answer module according to the analysis results.

[0004] Based on the description in the above documents, although the existing intelligent question-answering robot system based on online knowledge communities has made significant progress in many fields, there are still some problems that need to be solved urgently. In the process of processing user question language, the expressed questions are often not analyzed completely, so that the answers provided are not accurate enough, and the positive emotional value given to users when answering is relatively lacking. For this reason, the present invention provides an intelligent question-answering robot system based on online knowledge communities. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides an intelligent question-answering robot system based on an online knowledge community, which solves the problem that the existing intelligent question-answering robot system often fails to fully analyze the expressed questions during the process of language processing of user questions, so that the answers provided are not accurate enough and the positive emotional value given to users when answering is relatively lacking.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent question-answering robot system based on an online knowledge community, comprising:

[0007] User information module, where users complete the login of basic information through existing contact information;

[0008] A data interaction module, used to receive questions expressed by users in natural language and provide users with required information or answers;

[0009] Intelligent question-answering system, which is used to process natural language and provide feedback based on the matched answers after analysis, including:

[0010] The natural language processing module is connected to the data interaction module and is responsible for converting the user's natural language questions;

[0011] The question parsing and matching module is responsible for parsing user questions, evaluating users’ positive emotions and creative self-efficacy, and searching for matching information or answers in the knowledge base;

[0012] The information generation and return module is responsible for generating answers that match the user's questions and returning the answers to the data interaction module, while providing emotional support based on the user's emotional state;

[0013] The knowledge base management module is responsible for acquiring, organizing and updating knowledge data from online knowledge communities and storing the knowledge data in the knowledge base;

[0014] The learning and optimization module is connected to the knowledge base management module and is responsible for continuously optimizing the performance of the intelligent question-answering robot system based on user feedback and system usage.

[0015] Preferably, the knowledge in the knowledge base management module comes from multiple sources, including user-generated content, expert responses and published articles, and structured databases;

[0016] The knowledge base management module communicates with the online knowledge community through a local area network, regularly checks and updates the knowledge base, maintains the accuracy and timeliness of the content, and updates and optimizes the information and answers fed back based on the user's natural language expression questions.

[0017] Preferably, the specific processing operations of the natural language processing module are:

[0018] A1. After collecting the natural language questions expressed by the user, first complete the text cleaning operation of the natural language questions;

[0019] A2, extracting the conjunction lexical features from the cleaned natural language questions, marking the conjunction lexical features as N, and performing analysis based on the conjunction lexical features N to complete the primary and secondary sorting;

[0020] A3. Finally, the processed sequential conjunction lexical features N are aggregated to form a lexical feature dataset M.

[0021] Preferably, the text cleaning operation of the natural language question in A1 is:

[0022] a11. Remove special characters: Divide the natural language questions into multiple segments using a specific character set, and remove punctuation marks and special symbols in the text;

[0023] a12. Remove stop words: According to the stop word list, delete words that are common in the text but have no effect on the meaning of the sentence;

[0024] a13. Remove spaces and line breaks: Use string replacement to uniformly process redundant spaces and line breaks.

[0025] a14. Case standardization: Convert all text to a uniform case format.

[0026] Preferably, the extraction operation of conjunction vocabulary feature N in A2 is:

[0027] a21. Set a search function with adjustable matching word count and mark it as F(x);

[0028] a22, and the search function moves along the data sequence of the natural language question, starts from the beginning of the natural language question data through the search function F(x), and performs feature identification based on the number of matching words as two;

[0029] a23. Match the feature vocabulary of the feature identifier with the vocabulary set L in the knowledge base management module, and determine the conjunction vocabulary feature N according to the matching result.

[0030] Preferably, the specific operation steps of determining the conjunction vocabulary feature N based on the matching results in a22-a23 are:

[0031] C1. Performing a movement operation on the natural language question data based on the two matching word numbers by using the search function F(x), and completing the extraction operation of conjunction vocabulary features by using the feature extraction formula of the search function F(x);

[0032] C2, and the following situations exist after feature extraction and matching:

[0033] Case 1: After the feature word of the feature identifier is matched with the vocabulary set L in the knowledge base management module, the vocabulary set L contains vocabulary data containing the feature word or vocabulary data identical to the feature word;

[0034] Case 2: After the feature vocabulary of the feature identifier is matched with the vocabulary set L in the knowledge base management module, if there is no vocabulary data identical to the feature vocabulary in the vocabulary set L, then the vocabulary data related to the feature vocabulary is matched in the vocabulary set L;

[0035] C3, when the first condition is satisfied, by increasing the number of matching words of the search function F(x), the operations of C1-C2 are completed again, until the first condition is not satisfied, the vocabulary data marked by the previous search function F(x) under the current number of matching words is used as the conjunction vocabulary feature N, and the search function F(x) starting from the beginning of the natural language question data is ended, and a new search function F(x) is started with the last word of the search function F(x) under the current number of matching words as the starting point, and the operations of C1-C3 are repeated;

[0036] When situation 2 is satisfied, the vocabulary data of the search function F(x) under the current number of matching words is directly determined as the conjunction vocabulary feature N, and the search function F(x) starting from the beginning of the natural language question data is ended, and a new search function F(x) is started with the last word of the search function F(x) under the current number of matching words as the starting point, and the operations of C1-C3 are repeated.

[0037] Preferably, the feature extraction formula of the search function F(x) is: ; Where (1-n) and (n) are the weights of the search function F(x), and when the first condition is met, the weight of (1-n) decreases and the weight of (n) increases. On the contrary, when the second condition is met, the weight of (1-n) increases and the weight of (n) decreases. x is the lexical feature and the number of words in the vocabulary set L, N is the conjunction lexical feature, y is the lexical feature and the number of words related to the conjunction lexical feature in the vocabulary set L, and m is the number of words related to the conjunction lexical feature in the vocabulary set L. The vocabulary features and matching words added under the condition, and when the search function F(x) satisfies When the condition is met, continue to increase the number of words to search until it is met When , the judgment result is (x+m-1), that is, the vocabulary data marked by the previous search function F(x) of the search function F(x) under the current matching word number is used as the conjunction vocabulary feature N.

[0038] Preferably, the question analysis and matching module performs the following steps to analyze the user's question:

[0039] B1. After the conjunction lexical features N are determined, the lexical feature dataset M is introduced into the question parsing and matching module;

[0040] B2. At this time, the vocabulary range in the vocabulary feature data set M is screened one by one, first determining the direction of the problem elaboration, then determining the specific vocabulary features related to the elaboration direction, and finally determining the emotional vocabulary features expressed about the matter, and the combined vocabulary features are combined with the knowledge base management module and optimized;

[0041] B3. After the combination, the matching information or answers are found based on the knowledge base management module, and then the emotional language of the reply is added based on the emotional vocabulary characteristics.

[0042] Preferably, the optimization operation of the vocabulary features combined with the knowledge base management module in B2 is:

[0043] b21. First complete the matching operation of the exposition direction, and then match the conventional information reply in combination with the exposition direction and relevant specific matters;

[0044] b22. At this time, text data with sentiment labels are introduced as the optimization data set to fill in the sentiment sentences for the exposition direction and related specific matters matching the conventional information replies;

[0045] b23. After completing the preliminary emotional sentence feature filling, the emotional vocabulary features expressed about the matter are introduced and matched with the text data of the knowledge base and emotional tags to complete the filling and improvement operation of the reply sentence.

[0046] Preferably, the specific operation steps of the information generation and return module are:

[0047] D1. Generate the data results of the answers in the question analysis and matching module and send them to the user information module for review through the data interaction module;

[0048] D2. Take actions based on user feedback:

[0049] If the user is satisfied, no further questions will be asked and the conversation will end after the evaluation is made;

[0050] If the user continues to ask questions, repeat the operations of A1-A3 and B1-B3, and provide cognitive and emotional support based on the user's positive emotions and creativity self-efficacy status.

[0051] The present invention provides an intelligent question-answering robot system based on an online knowledge community. Compared with the prior art, it has the following beneficial effects:

[0052] (1) The intelligent question-answering robot system based on the online knowledge community, by setting up an intelligent question-answering system, adopts advanced natural language processing technology and artificial intelligence algorithms, realizes automatic parsing and understanding of user input, and completes the processing of natural language questions expressed by users through the natural language processing module, completes text cleaning and corresponding conjunction vocabulary feature extraction to improve the accuracy of understanding natural language, and sorts and combines conjunction vocabulary features to improve the accuracy of answer results.

[0053] (2) The intelligent question-answering robot system based on the online knowledge community, by setting up question parsing and matching modules, enables the intelligent question-answering robot to focus on establishing positive emotional connections with users. In the process of communicating with users, the system can show a friendly and respectful attitude, and combine text data with emotional labels as an optimization set to improve the situation value of the answer results and provide emotional support for users. This emotional support can not only help users relieve stress and enhance self-confidence, but also promote users to generate creative self-efficacy and positive emotions.

[0054] (3) The intelligent question-answering robot system based on the online knowledge community uses an intelligent method through the knowledge base management module to automatically acquire, organize and update knowledge from the online knowledge community, ensuring that the content of the knowledge base is always up-to-date. This improvement has greatly improved the accuracy and timeliness of the system's answers to questions, providing users with the latest knowledge information. In addition, through intelligent question-answering and emotional support, the present invention not only solves the user's instant query needs, but also stimulates the user's creativity and participation, promotes the internalization, socialization and innovation of knowledge, and thus significantly improves the activity of the online knowledge community and the speed of knowledge growth. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a functional block diagram of the intelligent question-answering robot system of the present invention;

[0056] Figure 2 is an operation flow chart of the natural language processing module of the present invention;

[0057] Figure 3 It is an operation flow chart of the conjunction vocabulary feature extraction of the present invention;

[0058] Figure 4 This is an operational flow chart of the problem analysis and matching module of the present invention. DETAILED DESCRIPTION

[0059] 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.

[0060] See also Figure 1-Figure 4 , the present invention provides three technical solutions:

[0061] Embodiment 1: An intelligent question-answering robot system based on an online knowledge community, comprising:

[0062] User information module, the user completes the login of basic information through the existing contact information. The user information module contains multiple information entry units, and K represents the number of information entry units, which are used to enable the user to complete the information filling and login operation;

[0063] A data interaction module, used to receive questions expressed by users in natural language and provide users with required information or answers;

[0064] Intelligent question-answering system, which is used to process natural language and provide feedback based on the matched answers after analysis, including:

[0065] The natural language processing module is connected to the data interaction module and is responsible for converting the user's natural language questions;

[0066] The question parsing and matching module is responsible for parsing user questions, evaluating users’ positive emotions and creative self-efficacy, and searching for matching information or answers in the knowledge base;

[0067] The information generation and return module is responsible for generating answers that match the user's questions and returning the answers to the data interaction module, while providing emotional support based on the user's emotional state;

[0068] The knowledge base management module is responsible for acquiring, organizing and updating knowledge data from online knowledge communities and storing the knowledge data in the knowledge base;

[0069] The learning and optimization module is connected to the knowledge base management module and is responsible for continuously optimizing the performance of the intelligent question-answering robot system based on user feedback and system usage.

[0070] Among them, by setting up an intelligent question-and-answer system and adopting advanced natural language processing technology and artificial intelligence algorithms, automatic parsing and understanding of user input are realized, and the natural language questions expressed by the user are processed through the natural language processing module, and text cleaning and corresponding conjunction vocabulary feature extraction are completed to improve the accuracy of understanding natural language, and conjunction vocabulary features are sorted and combined to improve the accuracy of the reply results.

[0071] In the embodiment of the present invention, the knowledge in the knowledge base management module comes from various sources, including user-generated content, expert replies and published articles, and structured databases;

[0072] The knowledge base management module communicates with the online knowledge community through the local area network, regularly checks and updates the knowledge base, maintains the accuracy and timeliness of the content, and updates and optimizes the information and answers fed back based on users' natural language expression of questions.

[0073] In the embodiment of the present invention, the specific processing operations of the natural language processing module are:

[0074] A1. After collecting the natural language questions expressed by the user, first complete the text cleaning operation of the natural language questions;

[0075] A2, extracting the conjunction lexical features from the cleaned natural language questions, marking the conjunction lexical features as N, and performing analysis based on the conjunction lexical features N to complete the primary and secondary sorting;

[0076] A3. Finally, the processed sequential conjunction lexical features N are aggregated to form a lexical feature dataset M.

[0077] In the embodiment of the present invention, the text cleaning operation of the natural language question in A1 is:

[0078] a11. Remove special characters: Divide the natural language questions into multiple segments using a specific character set, and remove punctuation marks and special symbols in the text;

[0079] a12. Remove stop words: According to the stop word list, delete words that are common in the text but have no effect on the meaning of the sentence;

[0080] a13. Remove spaces and line breaks: Use string replacement to uniformly process redundant spaces and line breaks.

[0081] a14. Case standardization: Convert all text to a uniform case format.

[0082] In the embodiment of the present invention, the extraction operation of conjunction vocabulary feature N in A2 is:

[0083] a21. Set a search function with adjustable matching word count and mark it as F(x);

[0084] a22, and the search function moves along the data sequence of the natural language question, starts from the beginning of the natural language question data through the search function F(x), and performs feature identification based on the number of matching words as two;

[0085] a23. Match the feature vocabulary of the feature identifier with the vocabulary set L in the knowledge base management module, and determine the conjunction vocabulary feature N according to the matching result.

[0086] In the embodiment of the present invention, the specific operation steps of determining the conjunction vocabulary feature N based on the matching results in a22-a23 are:

[0087] C1. Performing a movement operation on the natural language question data based on the two matching word numbers by using the search function F(x), and completing the extraction operation of conjunction vocabulary features by using the feature extraction formula of the search function F(x);

[0088] C2, and the following situations exist after feature extraction and matching:

[0089] Case 1: After the feature word of the feature identifier is matched with the vocabulary set L in the knowledge base management module, the vocabulary set L contains vocabulary data containing the feature word or vocabulary data identical to the feature word;

[0090] Case 2: After the feature vocabulary of the feature identifier is matched with the vocabulary set L in the knowledge base management module, if there is no vocabulary data identical to the feature vocabulary in the vocabulary set L, then the vocabulary data related to the feature vocabulary is matched in the vocabulary set L;

[0091] C3, when the first condition is satisfied, by increasing the number of matching words of the search function F(x), the operations of C1-C2 are completed again, until the first condition is not satisfied, the vocabulary data marked by the previous search function F(x) under the current number of matching words is used as the conjunction vocabulary feature N, and the search function F(x) starting from the beginning of the natural language question data is ended, and a new search function F(x) is started with the last word of the search function F(x) under the current number of matching words as the starting point, and the operations of C1-C3 are repeated;

[0092] When situation 2 is satisfied, the vocabulary data of the search function F(x) under the current number of matching words is directly determined as the conjunction vocabulary feature N, and the search function F(x) starting from the beginning of the natural language question data is ended, and a new search function F(x) is started with the last word of the search function F(x) under the current number of matching words as the starting point, and the operations of C1-C3 are repeated.

[0093] In the embodiment of the present invention, the feature extraction formula of the search function F(x) is: ; Wherein, (1-n) is the weight of (n) and the weight of the search function F(x). When the first condition is met, the weight of (1-n) decreases and the weight of (n) increases. On the contrary, when the second condition is met, the weight of (1-n) increases and the weight of (n) decreases. x is the vocabulary feature and the number of words in the vocabulary set L, N is the conjunction vocabulary feature, y is the vocabulary feature and the number of words related to the conjunction vocabulary feature in the vocabulary set L, and m is the number of words related to the conjunction vocabulary feature in the vocabulary set L. The vocabulary features and matching words added under the condition, and when the search function F(x) satisfies When the condition is met, continue to increase the number of words to search until it is met When , the judgment result is (x+m-1), that is, the vocabulary data marked by the previous search function F(x) of the search function F(x) under the current matching word number is used as the conjunction vocabulary feature N.

[0094] Among them, by setting up question analysis and matching modules, the intelligent question-answering robot focuses on establishing positive emotional connections with users. In the process of communicating with users, the system can show a friendly and respectful attitude, and combine text data with emotional labels as an optimization set to improve the situation value of the response results and provide emotional support for users. This emotional support can not only help users relieve stress and enhance self-confidence, but also promote users to generate creative self-efficacy and positive emotions.

[0095] In the embodiment of the present invention, the steps of the question analysis and matching module for analyzing the user question are as follows:

[0096] B1. After the conjunction lexical features N are determined, the lexical feature dataset M is introduced into the question parsing and matching module;

[0097] B2. At this time, the vocabulary range in the vocabulary feature data set M is screened one by one, first determining the direction of the problem elaboration, then determining the specific vocabulary features related to the elaboration direction, and finally determining the emotional vocabulary features expressed about the matter, and the combined vocabulary features are combined with the knowledge base management module and optimized;

[0098] B3. After the combination, the matching information or answers are found based on the knowledge base management module, and then the emotional language of the reply is added based on the emotional vocabulary characteristics.

[0099] In the embodiment of the present invention, the optimization operation of the vocabulary feature combined with the knowledge base management module of B2 is:

[0100] b21. First complete the matching operation of the exposition direction, and then match the conventional information reply in combination with the exposition direction and relevant specific matters;

[0101] b22. At this time, text data with sentiment labels are introduced as the optimization data set to fill in the sentiment sentences for the exposition direction and related specific matters matching the conventional information replies;

[0102] b23. After completing the preliminary emotional sentence feature filling, the emotional vocabulary features expressed about the matter are introduced and matched with the text data of the knowledge base and emotional tags to complete the filling and improvement operation of the reply sentence.

[0103] Among them, the knowledge base management module adopts intelligent methods to automatically acquire, organize and update knowledge from the online knowledge community, ensuring that the content of the knowledge base is always up to date. This progress has greatly improved the accuracy and timeliness of the system's answers to questions, and provided users with the latest knowledge information. Moreover, through intelligent question and answer and emotional support, the present invention not only solves the user's instant query needs, but also stimulates the user's creativity and participation, promotes the internalization, socialization and innovation of knowledge, thereby significantly improving the activity of the online knowledge community and the speed of knowledge growth.

[0104] In the embodiment of the present invention, the specific operation steps of the information generation and return module are:

[0105] D1. Generate the data results of the answers in the question analysis and matching module and send them to the user information module for review through the data interaction module;

[0106] D2. Take actions based on user feedback:

[0107] If the user is satisfied, no further questions will be asked and the conversation will end after the evaluation is made;

[0108] If the user continues to ask questions, repeat the operations of A1-A3 and B1-B3, and provide cognitive and emotional support based on the user's positive emotions and creativity self-efficacy status.

[0109] Embodiment 2, compared with Embodiment 1, is different in that: the question analysis and matching module is connected with the knowledge base management module and the natural language processing module, and is used to analyze user questions, evaluate the user's positive emotions (the degree of pleasure experienced by the user in the environment, including emotional states such as enthusiasm, high energy, mental alertness, and certainty), creativity self-efficacy (a self-belief or expectation that the user shows to himself when performing innovative activities), and search for matching information or answers in the knowledge base;

[0110] The information generation and feedback module provides certain cognitive and emotional support based on the user's positive emotions and creative self-efficacy status, adjusts the positive emotions or creative self-efficacy status to a higher level, and thus promotes the user to create new knowledge. This module analyzes the user's positive emotions and creative self-efficacy status, combines the user's questions and the knowledge or information searched from the knowledge base, as well as the information provided by the cognitive emotional support module, and finally integrates and feeds back to the user through the data interaction module. This module can also recommend users with similar experiences or common topics to existing interactive users through the data interaction module, thereby enhancing interaction and knowledge sharing within the community.

[0111] The information generation and return module generates answers that users can understand based on the matching results. The answers are not limited to text, but can also be in multiple formats such as charts and videos to enhance the expressiveness of the information and user experience. It can recommend similar users or related topics based on user interests and historical interactions, and promote communication and knowledge sharing within the community.

[0112] In addition, the learning and optimization module can analyze the results of interaction between users and the system, collect feedback information, use machine learning algorithms to analyze user feedback, identify deficiencies in system performance, and continuously adjust and optimize algorithms and databases based on the analysis results to improve system performance and user satisfaction. The intelligent question-and-answer robot system based on online knowledge communities can achieve efficient and accurate knowledge retrieval and information services, improve user experience, and promote knowledge sharing and community interaction.

[0113] Embodiment 3: The present invention also provides an application case in real life:

[0114] Real-life application case 1: There is an online health knowledge community where users share and exchange knowledge about healthy diet, exercise, disease prevention, etc. However, as the number of users increases, the information in the community becomes complicated, and users often find it difficult to quickly find the answers they need, or the answers they get are inaccurate, which leads to a decline in user experience and a gradual decrease in community activity;

[0115] In order to solve this problem, the online health knowledge community introduced the intelligent question-answering robot system of the present invention, which can not only quickly retrieve health knowledge related to the user's question from the knowledge base, but also provide personalized emotional support and suggestions based on the user's positive emotions and creativity self-efficacy status;

[0116] For example, a user asked a question in the community: "I always feel tired recently. Am I lacking some nutrients?" The intelligent question-answering robot system will first analyze the user's question, and then search for relevant information about fatigue and nutrients in the knowledge base. At the same time, the system will also evaluate the user's positive emotions and creativity self-efficacy status, and find that the user is somewhat anxious and uneasy;

[0117] Based on the above question language, the extracted vocabulary features are: "fatigue", "lack", "nutrients", "is it", "recently", "always", "feeling", and the elaboration direction is determined to be "feeling tired". Based on this elaboration direction, the user's main issue feature is "lack of nutrients". Based on the above, the user proposed the vocabulary features of "recently always" and "is it" for this type of question, so it is judged that the customer has frequent issues and has emotions of worry and doubt;

[0118] Therefore, in addition to providing professional answers about fatigue and nutrients, the system also sends a warm message through the user interaction module: "Please don't worry, fatigue is caused by many reasons. We can work together to find a solution that suits you. At the same time, it is also important to maintain a positive attitude and good living habits!" Such a reply not only solves the user's problem, but also gives the user emotional support, improving the user's satisfaction and trust.

[0119] In addition, the intelligent question-answering robot system will recommend other users who have similar experiences or common topics based on the user's interests and needs. These users can exchange experiences and insights with each other, further promoting the sharing and dissemination of knowledge. As time goes by, more and more users begin to actively use the intelligent question-answering robot system, and the interaction and communication within the community become more frequent and in-depth. Users not only find the answers they need here, but also make like-minded friends and jointly pursue a healthier and better lifestyle.

[0120] Through this real-life application example, it can be seen that the intelligent question-answering robot system of the present invention plays an important role in improving user experience, promoting knowledge sharing and enhancing community activity, and has brought tangible benefits to the development of online knowledge communities.

[0121] Application case 2 in real life: In an online knowledge community, a user asked the intelligent question-answering robot system: "I just found out and I'm very sad. In recent years, I often have negative emotions such as entanglement, worry, anxiety, depression, fear, etc. Is mental stress the cause of type 2 diabetes?" After searching, the intelligent question-answering robot system gave feedback: "Psychological stress, especially long-term stress, causes the adrenal glands to produce cortisol, which can cause: 1. The decomposition of adipose tissue and the delivery of fatty acids into blood vessels; 2. Inducing the liver to produce gluconeogenesis (converting fat into glucose) and deliver it to blood vessels; 3. Inducing the decomposition of glycogen in the liver and delivering the decomposed glucose into blood vessels; thereby causing the body's blood sugar to remain stable. At high levels, sustained high blood sugar leads to increased insulin resistance in the human body, that is, the content of insulin in the blood increases, but the efficacy decreases. At the same time, when people are under great pressure, they tend to drink alcohol and eat foods high in sugar, such as milk tea, ice cream, etc. (sugar can make people feel happy), and they also eat all kinds of snacks, etc. This vicious cycle makes diabetes worse. At the same time, feedback: "Adjust your mood. Various long-term mental stress will keep blood sugar at a high level. Type 2 diabetes is reversible. Believe in your body's strong self-healing ability and don't need any mental pressure; adjust your diet, communicate more with other patients, and adopt corresponding measures based on actual conditions. Believe that you will be able to reverse it."

[0122] The user continued to ask: "I feel much more relaxed. What measures should I take? Do I need to take medicine? How should I arrange my diet?" The intelligent question-answering robot system responded: "Current drugs can only control blood sugar levels, but cannot cure it. The first thing to do is to adjust your mind, supplemented by diet adjustment and exercise. Diabetes is reversible."

[0123] The user continued to ask: "I have some confidence now. I have decided to start with my mood and diet. So how do I arrange my diet?" The intelligent robot responded: "1. Adjust your mood. Various long-term mental stress will keep your blood sugar at a high level. Type 2 diabetes is reversible. Believe in your body's strong self-healing ability and there is no need for any mental stress. 2. Quit beverages (alcohol, carbonated drinks, fruit juice, milk tea). Eat fruit once every two days, and the amount at a time should not exceed the size of a fist. Adjust your diet structure, increase the proportion of meat, eggs, milk, and vegetables, eat a small handful of nuts every day, and relatively reduce the intake of "staple foods" such as rice and noodles. 3. Try to compress the time for eating within 8-10 hours a day, and drink water or tea at other times. 4. Get moving, do some exercises you like, strength or anaerobic training, which are more effective in burning sugar. 5. Have a good rest, you will get better soon. Taking medicine for a lifetime is not a treatment at all." User feedback: "OK, thank you very much".

[0124] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0125] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0126] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent question-answering robot system based on an online knowledge community, characterized in that: include: User information module, where users log in basic information using existing contact information and express questions in natural language; A data interaction module, used to receive questions expressed by users in natural language and provide users with required information or answers; Intelligent question-answering system, which is used to process natural language and provide feedback based on the matched answers after analysis, including: The natural language processing module is connected to the data interaction module, and extracts the vocabulary features from the user's natural language question recognition and transmits them to the question parsing and matching module for use; The question parsing and matching module is responsible for parsing user questions, evaluating the user's positive emotions and creative self-efficacy, searching for matching information or answers in the knowledge base, and transmitting the matching information or answers; The information generation and return module receives information or answer data, is responsible for generating answers that match the user's questions and returning the answers to the data interaction module, while providing emotional support based on the user's emotional state; The knowledge base management module is used to obtain, organize and update data from online knowledge communities and store knowledge data in the knowledge base for feature matching; The learning and optimization module is connected to the knowledge base management module and is responsible for continuously optimizing the performance of the intelligent question-answering robot system based on user feedback and system usage; The specific processing operations of the natural language processing module are: A1. After collecting the natural language questions expressed by the user, first complete the text cleaning operation of the natural language questions; A2, extracting conjunction lexical features from the cleaned natural language questions, marking the conjunction lexical features as N, and performing analysis based on the conjunction lexical features N to complete the primary and secondary sorting; A3. Finally, the processed sequential conjunction lexical features N are aggregated to form a lexical feature dataset M; The extraction operation of conjunction vocabulary feature N in A2 is: a21. Set a search function with adjustable matching word count and mark it as F(x); a22, and the search function moves along the data sequence of the natural language question, starts from the beginning of the natural language question data through the search function F(x), and performs feature identification based on the number of matching words as two; a23, matching the feature vocabulary of the feature identifier with the vocabulary set L in the knowledge base management module, and determining the conjunction vocabulary feature N according to the matching result; The specific operation steps of determining the conjunction vocabulary feature N based on the matching results in a22-a23 are: C1. Complete the extraction operation of conjunction vocabulary features by searching the feature extraction formula of function F(x); C2, and the following situations exist after feature extraction and matching: Case 1: After the feature word of the feature identifier is matched with the vocabulary set L in the knowledge base management module, the vocabulary set L contains vocabulary data containing the feature word or vocabulary data identical to the feature word; Case 2: After the feature vocabulary of the feature identifier is matched with the vocabulary set L in the knowledge base management module, if there is no vocabulary data identical to the feature vocabulary in the vocabulary set L, then the vocabulary data related to the feature vocabulary is matched in the vocabulary set L; C3, when the first condition is met, by increasing the number of matching words of the search function F(x), the operations of C1-C2 are completed again, until the first condition is not met, the vocabulary data of the previous search function F(x) under the current number of matching words is used as the conjunction vocabulary feature N, at which time the initial search function F(x) operation ends, and a new search function F(x) is started with the next word of the conjunction vocabulary feature N as the starting point for matching operation, and the operations of C1-C3 are repeated; When the second condition is satisfied, the vocabulary data of the search function F(x) under the current matching word count is directly determined as the conjunction vocabulary feature N, and the search function F(x) starting from the beginning of the natural language question data is terminated, and a new search function F(x) is started from the next word of the conjunction vocabulary feature N to perform matching operation, and the operations of C1-C3 are repeated; The feature extraction formula of the search function F(x) is: ； Where (1-n) and (n) are the weights of the search function F(x), and when the first condition is met, the weight of (1-n) decreases and the weight of (n) increases. On the contrary, when the second condition is met, the weight of (1-n) increases and the weight of (n) decreases. x is the lexical feature in the vocabulary set L, N is the conjunction lexical feature, y is the lexical feature related to the conjunction lexical feature in the vocabulary set L, and m is the lexical feature in the vocabulary set L. The vocabulary features added under the condition, and when the search function F(x) satisfies When the condition is met, continue to add vocabulary feature searches until the condition is met. When , the judgment result is (x+m-1), and the vocabulary data marked by the previous search function F(x) is used as the conjunction vocabulary feature N; The steps of the question analysis and matching module for analyzing user questions are as follows: B1. After the conjunction lexical features N are determined, the lexical feature dataset M is introduced into the question parsing and matching module; B2. At this time, the vocabulary range in the vocabulary feature data set M is screened one by one, first determining the direction of the problem elaboration, then determining the specific vocabulary features related to the elaboration direction, and finally determining the emotional vocabulary features expressed about the matter, and the combined vocabulary features are introduced into the knowledge base management module for update and optimization; The optimization operation of the vocabulary features combined with the knowledge base management module of B2 is: b21. First complete the matching operation of the exposition direction, and then match the conventional information reply in combination with the exposition direction and relevant specific matters; b22. At this time, text data with emotional tags are introduced as optimization data to fill in emotional sentences for regular information replies matching the exposition direction and related specific matters; b23. After completing the operation of filling in the emotional sentence features, the emotional vocabulary features expressed about the matter are introduced and matched with the text data of the emotional tags in the knowledge base to complete the filling and improvement operation of the reply sentence.

2. The intelligent question-answering robot system based on an online knowledge community according to claim 1, characterized in that: The knowledge in the knowledge base management module comes from multiple sources, including user-generated content, expert responses and published articles, and structured databases; The knowledge base management module communicates with the online knowledge community through a local area network, regularly checks and updates the knowledge base, and updates and optimizes the information and answers fed back based on the user's natural language expression questions.

3. The intelligent question-answering robot system based on an online knowledge community according to claim 1, characterized in that: The text cleaning operation for the natural language question in A1 is: a11. Remove special characters: Divide the natural language questions into multiple segments using a specific character set, and remove punctuation marks and special symbols in the text; a12. Remove stop words: According to the stop word list, delete words that are common in the text but have no effect on the meaning of the sentence; a13. Remove spaces and line breaks: Use string replacement to uniformly process redundant spaces and line breaks. a14. Case standardization: Convert all text to a uniform case format.

4. The intelligent question-answering robot system based on an online knowledge community according to claim 1, characterized in that: The specific operation steps of the information generation and return module are: D1. Generate the data results of the answers in the question analysis and matching module and send them to the user information module for review through the data interaction module; D2. Take actions based on user feedback: If the user is satisfied, no further questions will be asked and the conversation will end after the evaluation is made; If the user continues to ask questions, repeat the operations of A1-A3 and B1-B3, and provide cognitive and emotional support based on the user's positive emotions and creativity self-efficacy status.

Images