AI collaborative interaction system and device of cross-platform standardized database

Abstract

The invention discloses an AI collaborative interaction system and device of a cross-platform standardized database, and relates to the technical field of AI collaborative interaction systems, the AI collaborative interaction system comprises a data processing module, a standardization module, an interface module, a data management module, a prediction module and an interaction feedback module; the data processing module is used for collecting original data by adopting a distributed crawler technology to obtain a basic data set; the standardization module is used for carrying out standardization processing on the basic data set to obtain a standardized data packet; the interface module is used for performing cross-platform database access operation on the standardized data packet by adopting an abstraction layer design method to obtain a cross-platform query result; the cross-platform query result obtained by the interface module is analyzed and optimized by adopting the machine learning algorithm, the query performance is improved, the system can automatically identify and optimize the query path through the application of the machine learning algorithm, and the query efficiency is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of AI collaborative interaction systems, and in particular to an AI collaborative interaction system and device with a cross-platform standardized database. Background Art

[0002] The AI collaborative interaction system is an intelligent system that integrates multiple advanced technologies and modules. It aims to improve the efficiency and intelligence of data processing, management, and user interaction. Therefore, how to use advanced technical means to improve the intelligence and security of the AI collaborative interaction system has become one of the current urgent issues to be solved.

[0003] In the field of AI collaborative interaction systems, existing systems often face challenges such as low efficiency and low data quality when collecting data from multiple data sources. Data from different sources come in different formats, making data integration difficult. Existing user behavior analysis tools often lack support for deep learning, making it difficult to accurately predict users' future needs. At the same time, traditional systems only support a single form of user interaction and cannot provide a comprehensive user experience. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an AI collaborative interaction system with a cross-platform standardized database to solve the challenges of low efficiency and low data quality when collecting data from multiple data sources in existing systems. The data formats from different sources are different, which makes data integration difficult. In addition, existing user behavior analysis tools often lack support for deep learning, making it difficult to accurately predict users' future needs.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides an AI collaborative interaction system for a cross-platform standardized database, comprising:

[0008] Data processing module, standardization module, interface module, data management module, prediction module and interactive feedback module;

[0009] The data processing module is used to collect raw data using distributed crawler technology to obtain a basic data set;

[0010] The standardization module is used to perform standardization processing on the basic data set to obtain a standardized data packet;

[0011] The interface module is used to perform cross-platform database access operations on standardized data packets using an abstract layer design method to obtain cross-platform query results;

[0012] The data management module is used to analyze and optimize the cross-platform query results obtained by the interface module using a machine learning algorithm to obtain optimized query performance and status data;

[0013] The prediction module is used to analyze the optimized query performance and system status data using clustering algorithms, deep learning models, and causal reasoning knowledge graphs, extract user behavior patterns, and predict future demand to obtain user behavior prediction results;

[0014] The interactive feedback module is used to perform multimodal interactive design on user behavior prediction results by adopting natural language processing, speech recognition, image processing technology and a method combining causal reasoning knowledge graph with a large model, provide personalized service responses, and obtain personalized interactive feedback reports.

[0015] As a preferred solution of the AI collaborative interaction system of the cross-platform standardized database of the present invention, wherein: the distributed crawler technology is used to collect the original data to obtain the basic data set, and the specific steps are as follows:

[0016] Choose Apache Nutch as the distributed crawler tool;

[0017] Adjust Nutch configuration files according to the data source;

[0018] The data source includes a specific website or API;

[0019] Define the initial URL set and start the Nutch crawling process, collect raw data, and obtain the basic data set.

[0020] As a preferred solution of the AI collaborative interaction system of the cross-platform standardized database of the present invention, wherein: the basic data is standardized to obtain a standardized data packet, and the specific steps are:

[0021] Define data structure model according to business requirements;

[0022] Map and transform basic data according to the defined data structure model;

[0023] The data mapping formula is introduced to calculate the mapping accuracy of each data item. The expression is:

[0024]

[0025] Among them, M i represents the matching score of the i-th data item in the target data structure, T i is the total matching possibility of the corresponding data item;

[0026] Clean and validate the converted data to remove invalid or erroneous data;

[0027] The data processed in the above steps is encapsulated into a data packet in a standard format JSON.

[0028] As a preferred solution of the AI collaborative interaction system for a cross-platform standardized database of the present invention, wherein: the abstract layer design method is used to perform a cross-platform database access operation on the standardized data packet to obtain a cross-platform query result, and the specific steps are as follows:

[0029] Define an abstract data access layer that hides the specific implementation details of different database management systems DBMS;

[0030] For each supported DBMS, develop a specific adapter class to implement the IDataAccess interface;

[0031] Each adapter is responsible for converting the unified query request into the specific command of the corresponding DBMS and returning the result to the abstract layer.

[0032] Introducing connection pool technology to reduce the overhead of frequently establishing and closing database connections;

[0033] By pre-allocating a certain number of database connections and dynamically adjusting the connection pool size when needed, the expression is:

[0034]

[0035] Among them, the maximum capacity of the connection pool is C max , the current number of active connections is C active , U c is the connection pool usage;

[0036] Construct query statements based on business needs and call corresponding adapters through the abstract layer to execute query operations;

[0037] The query results are encapsulated into a data package in the standard format JSON.

[0038] As a preferred solution of the AI collaborative interaction system for the cross-platform standardized database of the present invention, wherein: the cross-platform query results obtained by the interface module are analyzed and optimized using a machine learning algorithm to obtain optimized query performance and status data, and the specific steps are as follows:

[0039] Preprocess the cross-platform query results obtained from the interface module;

[0040] The preprocessing steps include missing value filling and outlier detection and processing operations;

[0041] Based on historical query patterns, machine learning algorithms are applied to predict the fields most likely to be used in query conditions, and indexing strategies are dynamically adjusted to improve query efficiency.

[0042] Introduce the index efficiency evaluation formula to calculate the query performance score P under each index configuration i , the expression is:

[0043]

[0044] Among them, Q time_optimized Indicates the time required to execute the query under the current index configuration, Q time_original Indicates the time required to execute the same query under unoptimized index configuration;

[0045] Generate the optimal query execution plan by analyzing the current database status and query requirements;

[0046] Use the decision tree model to select the best query path. The expression is:

[0047] BestPath=DTree(QueryParams);

[0048] Among them, QueryParams represents the parameter set in the query request, and BestPath represents the best query execution path recommended by the decision tree model;

[0049] Deploy monitoring tools to collect system performance indicators and make improvement suggestions based on the system performance indicator data;

[0050] Use the linear regression model to predict future system load, the expression is:

[0051] L t =α+β1T t +β2R t +∈;

[0052] Among them, L t represents the predicted system load, T t represents the CPU utilization at time t, R t represents the average response time at time t, α is the intercept term, β1 is the coefficient of CPU utilization, β2 is the coefficient of response time, and ∈ represents random error.

[0053] As a preferred solution of the AI collaborative interaction system for the cross-platform standardized database of the present invention, the method for using a clustering algorithm, a deep learning model, and a causal reasoning knowledge graph to analyze the optimized query performance and system status data, extract user behavior patterns and predict future demand, and obtain user behavior prediction results, specifically comprises the following steps:

[0054] Obtain optimized query performance and system status data from the data management module;

[0055] The data includes user query records, response time and CPU utilization;

[0056] Apply K-means clustering algorithm to cluster user behavior data;

[0057] Select the number of clusters and determine the optimal K value using the elbow method, which is expressed as:

[0058]

[0059] Where WCSS is the sum of all the squares within the cluster, C k represents all points in the kth cluster, x represents a point in the cluster, μ k represents the center point of the kth cluster;

[0060] Based on the clustering results, the long short-term memory network model LSTM is used to predict users' future needs;

[0061] The LSTM model is trained using the causal-aware loss function, which is expressed as:

[0062]

[0063] in, is the predicted demand, y t Represents the user's past behavior data;

[0064] Combining the clustering results with the prediction output of the LSTM model, the causal reasoning knowledge graph is used to verify the prediction results and extract the user's behavior pattern;

[0065] After the above steps, the user behavior prediction result is obtained.

[0066] As a preferred solution of the cross-platform standardized database AI collaborative interaction system described in the present invention, wherein: the method of combining natural language processing, speech recognition, image processing technology, and causal reasoning knowledge graph with a large model is used to perform multimodal interaction design on user behavior prediction results, provide personalized service responses, and obtain personalized interaction feedback reports. The specific steps are as follows:

[0067] Apply NLP technology to analyze user queries and feedback to generate personalized responses or suggestions;

[0068] Use the BERT model to understand user intent and generate appropriate responses;

[0069] When generating responses, the causal reasoning knowledge graph is queried in real time to verify the correctness of the answer, reducing the possibility of "hallucination" output;

[0070] For scenarios requiring voice interaction, speech recognition technology is used to convert the user's voice commands into text, and speech synthesis technology is used to convert the system's responses into voice output;

[0071] Use Google's Speech-to-Text API for speech recognition and Text-to-Speech API for speech synthesis. The expression is:

[0072] RT=STT model (UserVoice);

[0073] SSp=TTS model (SR);

[0074] Among them, RT represents the text recognized from the user's voice, UserVoice represents the user's voice input, SSp represents the result of converting the system response into voice output, and SR represents the response text generated by the system;

[0075] For scenarios involving image data, use image processing technology to generate visual feedback reports;

[0076] Integrate the results generated by the above modules to form a multimodal interactive feedback report;

[0077] Use the causal reasoning knowledge graph as a "verifier" to ensure the accuracy and consistency of feedback information;

[0078] The large model acts as an "enhancer" and will feed back the newly discovered rules after expert review to expand the coverage of the knowledge graph.

[0079] In a second aspect, the present invention provides an AI collaborative interaction device for a cross-platform standardized database, and when the processor executes the computer program, it implements the steps of the AI collaborative interaction system for a cross-platform standardized database as described in any one of claims 1 to 7.

[0080] The beneficial effects of the present invention are as follows: by using a machine learning algorithm to analyze and optimize the cross-platform query results obtained by the interface module, query performance is improved. Through the application of the machine learning algorithm, the system can automatically identify and optimize the query path, significantly improving query efficiency. In combination with the decision tree model to select the optimal query path, the query response speed is further improved. At the same time, by real-time monitoring of system performance indicators and proposing improvement suggestions, the continuous optimization and efficient operation of the system are ensured. By using an abstract layer design method to perform cross-platform database access operations on standardized data packets, efficient database queries are achieved. By introducing connection pool technology and abstract layer design, the system can effectively manage database connections and reduce resource consumption. The design of the abstract layer enables the system to seamlessly switch between different database management systems, enhancing the flexibility and portability of the system, and achieving the purpose of improving query performance and system stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0082] Figure 1 Schematic diagram of the AI collaborative interaction system for the cross-platform standardized database in Example 1. DETAILED DESCRIPTION

[0083] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0084] Reference Figure 1 , provides an AI collaborative interaction system with a cross-platform standardized database, including:

[0085] Data processing module, standardization module, interface module, data management module, prediction module and interactive feedback module

[0086] The data processing module is used to collect raw data using distributed crawler technology to obtain a basic data set;

[0087] Going a step further, Apache Nutch was chosen as the distributed crawler tool;

[0088] Adjust Nutch configuration files according to the data source;

[0089] Data sources include specific websites or APIs;

[0090] Define the initial URL set and start the Nutch crawling process, collect raw data, and obtain the basic data set;

[0091] It should be noted that Apache Nutch was chosen as the distributed crawler tool not only because of its strong scalability and flexibility, but also because it supports multiple protocols and data sources, which enables the system to efficiently collect data from a wide range of network resources. In addition, by adjusting Nutch's configuration file to adapt to different data sources, the accuracy and efficiency of the data collection process can be ensured, laying a solid foundation for subsequent data processing and analysis.

[0092] The standardization module is used to standardize the basic data set to obtain a standardized data package;

[0093] Going further, define the data structure model according to business requirements;

[0094] Map and transform basic data according to the defined data structure model;

[0095] The data mapping formula is introduced to calculate the mapping accuracy of each data item. The expression is:

[0096]

[0097] Among them, M i represents the matching score of the i-th data item in the target data structure, T i is the total matching possibility of the corresponding data item;

[0098] Clean and validate the converted data to remove invalid or erroneous data;

[0099] Encapsulate the data processed in the above steps into a standard format data packet JSON;

[0100] It should be noted that the introduction of data mapping formulas to calculate the mapping accuracy of each data item can not only evaluate the quality of data conversion, but also help identify and correct problems that may arise during the conversion process. By cleaning and verifying the converted data and removing invalid or erroneous data, the reliability and consistency of the data set are further improved, ensuring the smooth progress of subsequent processing steps.

[0101] The interface module is used to perform cross-platform database access operations on standardized data packets using an abstract layer design method to obtain cross-platform query results;

[0102] Furthermore, an abstract data access layer is defined, which hides the specific implementation details of different database management systems DBMS;

[0103] For each supported DBMS, develop a specific adapter class to implement the IDataAccess interface;

[0104] Each adapter is responsible for converting the unified query request into the specific command of the corresponding DBMS and returning the result to the abstract layer.

[0105] Introducing connection pool technology to reduce the overhead of frequently establishing and closing database connections;

[0106] By pre-allocating a certain number of database connections and dynamically adjusting the connection pool size when needed, the expression is:

[0107]

[0108] Among them, the maximum capacity of the connection pool is C max , the current number of active connections is C active , U c is the connection pool usage;

[0109] Construct query statements based on business needs and call corresponding adapters through the abstract layer to execute query operations;

[0110] The query results are encapsulated into a standard format data packet JSON;

[0111] It should be noted that the use of connection pool technology and dynamic adjustment of the connection pool size through formulas can effectively reduce the overhead of establishing and closing database connections, thereby improving the system's response speed and stability. The abstract layer design method hides the specific implementation details of different DBMSs, making the system more flexible and portable, and able to easily cope with diverse database environments.

[0112] The data management module is used to analyze and optimize the cross-platform query results obtained by the interface module using machine learning algorithms to obtain optimized query performance and status data;

[0113] Furthermore, the cross-platform query results obtained from the interface module are pre-processed;

[0114] The preprocessing steps include missing value filling and outlier detection and processing operations;

[0115] Based on historical query patterns, machine learning algorithms are applied to predict the fields most likely to be used in query conditions, and indexing strategies are dynamically adjusted to improve query efficiency.

[0116] Introduce the index efficiency evaluation formula to calculate the query performance score P under each index configuration i , the expression is:

[0117]

[0118] Among them, Q time_optimized Indicates the time required to execute the query under the current index configuration, Q time_original Indicates the time required to execute the same query under unoptimized index configuration;

[0119] Generate the optimal query execution plan by analyzing the current database status and query requirements;

[0120] Use the decision tree model to select the best query path. The expression is:

[0121] BestPath=DTree(QueryParams);

[0122] Among them, QueryParams represents the parameter set in the query request, and BestPath represents the best query execution path recommended by the decision tree model;

[0123] Deploy monitoring tools to collect system performance indicators and make improvement suggestions based on the system performance indicator data;

[0124] Use the linear regression model to predict future system load, the expression is:

[0125] L t =α+β1T t +β2R t +∈;

[0126] Among them, L t represents the predicted system load, T t represents the CPU utilization at time t, R t represents the average response time at time t, α is the intercept term, β1 is the coefficient of CPU utilization, β2 is the coefficient of response time, and ∈ represents random error;

[0127] It should be noted that the index efficiency evaluation formula is used to measure the effectiveness of index optimization, help identify query paths that need further optimization, use the decision tree model to generate the optimal query execution plan, and combine it with the linear regression model to predict future system load, so that the system can still maintain an efficient operating state under high load conditions. The application of real-time performance monitoring tools ensures that the system can continuously optimize itself and provide stable services.

[0128] The prediction module uses clustering algorithms, deep learning models, and causal reasoning knowledge graphs to analyze optimized query performance and system status data, extract user behavior patterns, and predict future demand to obtain user behavior prediction results.

[0129] Furthermore, the optimized query performance and system status data are obtained from the data management module;

[0130] The data includes user query records, response time, and CPU utilization;

[0131] Apply K-means clustering algorithm to cluster user behavior data;

[0132] Select the number of clusters and determine the optimal K value using the elbow method, which is expressed as:

[0133]

[0134] Where WCSS is the sum of all the squares within the cluster, C k represents all points in the kth cluster, x represents a point in the cluster, μ k represents the center point of the kth cluster;

[0135] Based on the clustering results, the long short-term memory network model LSTM is used to predict users' future needs;

[0136] The LSTM model is trained using the causal-aware loss function, which is expressed as:

[0137]

[0138] in, is the predicted demand, y t Represents the user's past behavior data;

[0139] Combining the clustering results with the prediction output of the LSTM model, the causal reasoning knowledge graph is used to verify the prediction results and extract the user's behavior pattern;

[0140] After the above steps, the user behavior prediction results are obtained;

[0141] It should be noted that the selection of the K-means clustering algorithm and the determination of its optimal K value help to accurately divide user groups and understand the behavioral patterns of different users. Based on these patterns, the LSTM model can more accurately predict future needs and improve the accuracy of personalized services. The method of combining deep learning with traditional machine learning not only improves the accuracy of predictions, but also provides users with a more personalized experience.

[0142] The interactive feedback module is used to design multimodal interactions based on user behavior prediction results using natural language processing, speech recognition, image processing technology, and a combination of causal reasoning knowledge graphs and large models, providing personalized service responses and generating personalized interactive feedback reports.

[0143] Furthermore, NLP technology is applied to analyze user queries and feedback to generate personalized responses or suggestions;

[0144] Use the BERT model to understand user intent and generate appropriate responses;

[0145] When generating responses, the causal reasoning knowledge graph is queried in real time to verify the correctness of the answer, reducing the possibility of "hallucination" output;

[0146] For scenarios requiring voice interaction, speech recognition technology is used to convert the user's voice commands into text, and speech synthesis technology is used to convert the system's responses into voice output;

[0147] Use Google's Speech-to-Text API for speech recognition and Text-to-Speech API for speech synthesis. The expression is:

[0148] RT=STT model (UserVoice);

[0149] SSp=TTS model (SR);

[0150] Among them, RT represents the text recognized from the user's voice, UserVoice represents the user's voice input, SSp represents the result of converting the system response into voice output, and SR represents the response text generated by the system;

[0151] For scenarios involving image data, use image processing technology to generate visual feedback reports;

[0152] Integrate the results generated by the above modules to form a multimodal interactive feedback report;

[0153] Use the causal reasoning knowledge graph as a "verifier" to ensure the accuracy and consistency of feedback information;

[0154] The large model acts as an "enhancer," feeding back the discovered new patterns after expert review to expand the coverage of the knowledge graph;

[0155] It should be noted that by combining natural language processing, speech recognition and image processing technologies, the interactive feedback module can generate multimodal feedback reports based on user preferences and behavior patterns, understand user intentions through the BERT model and generate corresponding replies, or use Google's Speech-to-Text API and Text-to-Speech API to achieve voice interaction. This method not only enhances the user experience, but also enables the system to better meet diverse needs and provide a more intuitive and convenient service method.

[0156] This embodiment also provides an AI collaborative interaction device for a cross-platform standardized database, which is a computer device with a storage medium, and the storage medium is loaded with a computer program.

[0157] In summary, the present invention improves query performance by using a machine learning algorithm to analyze and optimize the cross-platform query results obtained by the interface module. Through the application of the machine learning algorithm, the system can automatically identify and optimize the query path, significantly improving query efficiency. In combination with the decision tree model, the optimal query path is selected, further improving the query response speed. At the same time, by real-time monitoring of system performance indicators and proposing improvement suggestions, the continuous optimization and efficient operation of the system are ensured. By adopting an abstract layer design method to perform cross-platform database access operations on standardized data packets, efficient database queries are achieved. By introducing connection pool technology and abstract layer design, the system can effectively manage database connections and reduce resource consumption. The design of the abstract layer enables the system to seamlessly switch between different database management systems, enhancing the flexibility and portability of the system, and achieving the purpose of improving query performance and system stability.

Claims

1. An AI collaborative interaction system with a cross-platform standardized database, characterized by: include: Data processing module, standardization module, interface module, data management module, prediction module and interactive feedback module; The data processing module is used to collect raw data using distributed crawler technology to obtain a basic data set; The standardization module is used to perform standardization processing on the basic data set to obtain a standardized data packet; The interface module is used to perform cross-platform database access operations on standardized data packets using an abstract layer design method to obtain cross-platform query results; The data management module is used to analyze and optimize the cross-platform query results obtained by the interface module using a machine learning algorithm to obtain optimized query performance and status data; The prediction module is used to analyze the optimized query performance and system status data using clustering algorithms, deep learning models, and causal reasoning knowledge graphs, extract user behavior patterns, and predict future demand to obtain user behavior prediction results; The interactive feedback module is used to perform multimodal interactive design on user behavior prediction results by adopting natural language processing, speech recognition, image processing technology and a method combining causal reasoning knowledge graph with a large model, provide personalized service responses, and obtain personalized interactive feedback reports.

2. The cross-platform standardized database AI collaborative interaction system according to claim 1, characterized in that: The distributed crawler technology is used to collect the original data to obtain the basic data set. The specific steps are as follows: Choose Apache Nutch as the distributed crawler tool; Adjust Nutch configuration files according to the data source; The data source includes a specific website or API; Define the initial URL set and start the Nutch crawling process, collect raw data, and obtain the basic data set.

3. The cross-platform standardized database AI collaborative interaction system according to claim 2, characterized in that: The basic data is standardized to obtain a standardized data packet, and the specific steps are as follows: Define data structure model according to business requirements; Map and transform basic data according to the defined data structure model; The data mapping formula is introduced to calculate the mapping accuracy of each data item. The expression is: Among them, M i represents the matching score of the i-th data item in the target data structure, T i is the total matching possibility of the corresponding data item; Clean and validate the converted data to remove invalid or erroneous data; The data processed in the above steps is encapsulated into a data packet in a standard format JSON.

4. The cross-platform standardized database AI collaborative interaction system according to claim 3, characterized in that: The abstract layer design method is used to perform cross-platform database access operations on standardized data packets to obtain cross-platform query results. The specific steps are as follows: Define an abstract data access layer that hides the specific implementation details of different database management systems DBMS; For each supported DBMS, develop a specific adapter class to implement the IDataAccess interface; Each adapter is responsible for converting unified query requests into specific commands of the corresponding DBMS and returning the results to the abstraction layer; Introducing connection pool technology to reduce the overhead of frequently establishing and closing database connections; By pre-allocating a certain number of database connections and dynamically adjusting the connection pool size when needed, the expression is: Among them, the maximum capacity of the connection pool is C max , the current number of active connections is C active , U c is the connection pool usage; Construct query statements based on business needs and call corresponding adapters through the abstract layer to execute query operations; The query results are encapsulated into a data package in the standard format JSON.

5. The cross-platform standardized database AI collaborative interaction system according to claim 4, characterized in that: The machine learning algorithm is used to analyze and optimize the cross-platform query results obtained by the interface module to obtain optimized query performance and status data. The specific steps are as follows: Preprocess the cross-platform query results obtained from the interface module; The preprocessing steps include missing value filling and outlier detection and processing operations; Based on historical query patterns, machine learning algorithms are applied to predict the fields most likely to be used in query conditions, and indexing strategies are dynamically adjusted to improve query efficiency. Introduce the index efficiency evaluation formula to calculate the query performance score P under each index configuration i , the expression is: Among them, Q time_optimized Indicates the time required to execute the query under the current index configuration, Q time_original Indicates the time required to execute the same query under unoptimized index configuration; Generate the optimal query execution plan by analyzing the current database status and query requirements; Use the decision tree model to select the best query path. The expression is: BestPath=DTree(QueryParams); Among them, QueryParams represents the parameter set in the query request, and BestPath represents the best query execution path recommended by the decision tree model; Deploy monitoring tools to collect system performance indicators and make improvement suggestions based on the system performance indicator data; Use the linear regression model to predict future system load, the expression is: L t =α+β1T t +β2R t +∈; Among them, L t represents the predicted system load, T t represents the CPU utilization at time t, R t represents the average response time at time t, α is the intercept term, β1 is the coefficient of CPU utilization, β2 is the coefficient of response time, and ∈ represents random error.

6. The cross-platform standardized database AI collaborative interaction system according to claim 5, characterized in that: The method is used to analyze the optimized query performance and system status data using clustering algorithms, deep learning models, and causal reasoning knowledge graphs, extract user behavior patterns, and predict future demand to obtain user behavior prediction results. The specific steps are as follows: Obtain optimized query performance and system status data from the data management module; The data includes user query records, response time and CPU utilization; Apply K-means clustering algorithm to cluster user behavior data; Select the number of clusters and determine the optimal K value using the elbow method, which is expressed as: Where WCSS is the sum of all the squares within the cluster, C k represents all points in the kth cluster, x represents a point in the cluster, μ k represents the center point of the kth cluster; Based on the clustering results, the long short-term memory network model LSTM is used to predict users' future needs; The LSTM model is trained using the causal-aware loss function, which is expressed as: in, is the predicted demand, y t Represents the user's past behavior data; Combining the clustering results with the prediction output of the LSTM model, the causal reasoning knowledge graph is used to verify the prediction results and extract the user's behavior pattern; After the above steps, the user behavior prediction result is obtained.

7. The cross-platform standardized database AI collaborative interaction system according to claim 6, characterized in that: The method of combining natural language processing, speech recognition, image processing technology, causal reasoning knowledge graph and large model is used to perform multimodal interaction design on user behavior prediction results, provide personalized service response, and obtain personalized interaction feedback report. The specific steps are as follows: Apply NLP technology to analyze user queries and feedback to generate personalized responses or suggestions; Use the BERT model to understand user intent and generate appropriate responses; When generating responses, the causal reasoning knowledge graph is queried in real time to verify the correctness of the answer, reducing the possibility of "hallucination" output; For scenarios requiring voice interaction, speech recognition technology is used to convert the user's voice commands into text, and speech synthesis technology is used to convert the system's responses into voice output; Use Google's Speech-to-Text API for speech recognition and Text-to-Speech API for speech synthesis. The expression is: RT=STT model (UserVoice); SSp=TTS model (SR); Among them, RT represents the text recognized from the user's voice, UserVoice represents the user's voice input, SSp represents the result of converting the system response into voice output, and SR represents the response text generated by the system; For scenarios involving image data, use image processing technology to generate visual feedback reports; Integrate the results generated by the above modules to form a multimodal interactive feedback report; Use the causal reasoning knowledge graph as a "verifier" to ensure the accuracy and consistency of feedback information; The large model acts as an "enhancer" and will feed back the newly discovered rules after expert review to expand the coverage of the knowledge graph.

8. An AI collaborative interaction device for a cross-platform standardized database, comprising the AI collaborative interaction system for a cross-platform standardized database according to any one of claims 1 to 7, characterized in that: The device is a computer device having a storage medium, and the storage medium is loaded with a computer program.

Images