Data interaction system based on artificial intelligence

Through the data interaction method based on artificial intelligence, the problems of low data interaction efficiency, poor accuracy, poor user experience and insufficient data security in the existing technology are solved, and efficient, accurate, intelligent and secure data interaction is achieved, which improves the System performance and user experience.

CN119025539AInactive Publication Date: 2024-11-26TIANJIN JUNRUI TECHNOLOGY CO LTD

Patent Information

Application Number
CN202411048136.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-11-26
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing data interaction technology has many shortcomings in data collection, preprocessing, model construction, user interaction, data management, intelligent screening, parameter selection, system structure and security, resulting in low data interaction efficiency, poor accuracy, and poor user experience. Excellent, system performance degraded and data security insufficient.

Method used

Adopt artificial intelligence-based data interaction methods, including data acquisition module, data preprocessing module, artificial intelligence model module, data interaction module, cache module, user interface module, data management optimization module, intelligent filtering module, parameter selection module and security and The permission management module realizes efficient, accurate and secure data interaction through natural language processing, distributed computing, intelligent data screening, encryption technology and data desensitization.

Benefits of technology

It significantly improves the efficiency and accuracy of data interaction, enhances the adaptability and intelligence of the system, improves the user experience, solves the problem of data overload, guarantees the security and privacy of data, and improves the accuracy and model of the algorithm. performance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119025539A_ABST
    Figure CN119025539A_ABST
Patent Text Reader

Abstract

The invention relates to a data interaction system based on artificial intelligence. The data interaction system comprises a data acquisition module, a data preprocessing module, an artificial intelligence model module, a data interaction module, a cache module, a user interface module, a data management optimization module, an intelligent screening module, a parameter selection module and a security and authority management module. The system can collect original data from various data sources, perform cleaning, conversion and normalization processing, perform analysis through an artificial intelligence model, realize efficient interaction by using a data interaction module, improve response speed by combining with a cache module, provide a friendly user interface, perform data management optimization, intelligent screening and parameter selection, and perform data management and optimization. And meanwhile, the security and authority management of the data are ensured, and comprehensive, efficient, accurate and secure data interaction services are provided for the user.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data interaction technology, and in particular to an artificial intelligence-based data interaction method and system. Background Art

[0002] In today's digital age, data has become an important asset for enterprises and organizations. Efficient data interaction and accurate analysis are crucial for decision-making, business optimization, and user experience improvement. However, existing data interaction technologies face many severe challenges and shortcomings in practical applications.

[0003] Traditional data interaction methods usually rely on manually written rules and fixed processes to process data, which is not only inefficient but also difficult to cope with the diversity and complexity of data. When faced with massive amounts of data, data collection is often incomplete and inaccurate, resulting in deviations in subsequent analysis and application. At the same time, there are also many problems in the data preprocessing process, such as incomplete data cleaning, inability to effectively remove noise and erroneous data, and inflexible data conversion and normalization operations that cannot adapt to the needs of different data sources and formats.

[0004] In terms of model building and analysis, the models used by traditional methods are usually based on fixed parameters and structures, and cannot be adaptively adjusted according to the dynamic changes of data and real-time feedback of user behavior, resulting in limited accuracy and generalization of the models. Moreover, these models are computationally inefficient when processing large-scale data and are difficult to meet real-time requirements.

[0005] In terms of user interaction, most existing systems are unable to understand natural language input, and users need to use complex operations and specific instructions to obtain the required data, which greatly increases the user's usage threshold and operating costs, and seriously affects the user experience and the efficiency of data interaction.

[0006] In terms of data management and optimization, traditional technologies lack effective data classification, compression and archiving strategies, resulting in waste of data storage space, increased processing complexity, serious data overload problems and a sharp decline in system performance.

[0007] In terms of intelligent screening and recommendation, existing algorithms often cannot accurately capture users' interests and preferences, and the relevance and personalization of recommendation results are low, which cannot meet users' actual needs.

[0008] In addition, in terms of parameter selection, there is a lack of scientific and reasonable methods and tools, and people often rely on manual experience and repeated experiments, which is not only inefficient but also difficult to find the optimal combination of parameters, which seriously affects the accuracy of the algorithm and the performance of the model.

[0009] In terms of system structure and security, the existing system modules are not tightly connected and the data flow is not clear, resulting in low data transmission efficiency and prone to security vulnerabilities. The security and privacy protection of data during transmission and storage is insufficient, the user's access rights are not flexible enough, and data security is difficult to be effectively guaranteed.

[0010] To sum up, the existing data interaction technology has many shortcomings and obstacles in data collection, preprocessing, model building, user interaction, data management, intelligent screening, parameter selection, system structure and security, which seriously restrict the value mining and application effect of data. There is an urgent need for an innovative, artificial intelligence-based data interaction method and system to solve these problems and achieve efficient, accurate, intelligent and secure data interaction. Summary of the invention

[0011] The purpose of the present invention is to provide a data interaction method and system based on artificial intelligence, to solve the deficiencies and defects of existing data interaction technologies in the whole process of data collection and preprocessing, model building, user interaction, data management, intelligent screening and recommendation, parameter selection, system structure and data security, and to achieve efficient, accurate, intelligent and secure data interaction, including:

[0012] The data acquisition module is used to collect raw data from various data sources such as databases, file systems, sensors, etc., and pass the collected data to the data preprocessing module;

[0013] The data preprocessing module is connected to the data acquisition module and is responsible for cleaning the received raw data, removing noise, duplication and erroneous data, converting data, unifying data in different formats into a standard format, and performing data normalization operations to make the data consistent and comparable. The processed data is transmitted to the artificial intelligence model module;

[0014] The artificial intelligence model module is connected to the data preprocessing module;

[0015] The data interaction module is connected to the artificial intelligence model module, and uses natural language processing technology to understand the natural language needs of users and convert them into executable query and operation instructions. It adopts a distributed computing framework to distribute data processing tasks to multiple computing nodes for parallel execution, dynamically adjusts the allocation of computing resources through load balancing algorithms, and uses intelligent data screening algorithms to predict data that users may be interested in based on user needs and historical behaviors, and sends relevant instructions to the cache module in advance for caching;

[0016] A cache module, connected to the data interaction module, is used to store the predicted data that the user may be interested in, so as to improve the response speed of data interaction;

[0017] The user interface module is connected to the data interaction module to provide users with a friendly operation interface, making it convenient for users to input requirements and obtain results;

[0018] A data management optimization module is connected to the data preprocessing module and the data interaction module respectively;

[0019] Intelligent screening module, connected with the data interaction module;

[0020] A parameter selection module connected to the artificial intelligence model module;

[0021] The security and permission management module is connected to the above modules and adopts encryption technology and data desensitization methods to ensure the security and privacy of data during transmission and storage. At the same time, it supports users to flexibly set data access rights to ensure user data security.

[0022] Furthermore, the artificial intelligence model module can calculate the gradient L according to formula 1 using the back propagation algorithm based on data features and user behavior, and adjust the weight of the neural network according to the gradient L. The parameter adjustment method is based on the gradient descent method, and the step size of the parameter update is controlled by setting a suitable learning rate;

[0023]

[0024] Where N is the number of samples, y i is the true value, is the predicted value.

[0025] Furthermore, the data management optimization module uses data classification strategy to classify data, adjusts parameters by changing the initial value of the cluster center and the number of iterations, and optimizes the clustering effect. The classification formula is implemented using Formula 2:

[0026]

[0027] Among them, x i and x j are two data points and n is the number of features.

[0028] Furthermore, the intelligent screening module uses a collaborative filtering algorithm to predict the data that the user is interested in based on the user's historical behavior and the preferences of similar users, and optimizes the parameters by adjusting the weights or thresholds of the commonly evaluated items. The similarity calculation formula is shown in Formula 3:

[0029]

[0030] Among them, u and v are two users, I uv is the set of items that users u and v have evaluated together, r ui and rvi are the user and the rating of the item respectively.

[0031] Furthermore, the data interaction module can dynamically optimize distributed computing and load balancing strategies according to real-time data traffic and the load conditions of computing nodes.

[0032] Furthermore, the data management optimization module can automatically adjust the strategies of data classification, compression and archiving according to the growth and change trend of data.

[0033] Furthermore, the intelligent screening module can dynamically update the user interest prediction model by combining real-time user behavior and new user preference data.

[0034] Furthermore, the parameter selection module can evaluate the impact of parameter selection on model performance in real time during model training, and dynamically adjust the parameter value range, for example, setting the initial value range of the learning rate to [0.001, 0.1] based on experience, and then further screening the optimal value within this range through cross-validation.

[0035] Furthermore, the user interface module can adaptively adjust the interface layout and function display according to the user's usage habits and operating preferences.

[0036] Furthermore, data transmission and interaction between modules follow efficient, stable and secure communication protocols.

[0037] In summary, the beneficial technical effects of this application are:

[0038] It significantly improves the efficiency and accuracy of data interaction, can quickly and accurately process large-scale and complex data, and provide high-quality analysis and prediction results.

[0039] Furthermore, the adaptability and intelligence of the system are greatly enhanced, and it can automatically optimize models and parameters according to real-time changing data and user behavior, ensuring that the system always maintains good performance.

[0040] Furthermore, the user experience has been significantly improved. Through natural language processing and personalized interactive services, it is more in line with user needs, allowing users to obtain the required data and services more conveniently and efficiently.

[0041] Furthermore, the data overload problem was effectively solved, and the system performance and response speed were greatly improved through data management and optimization and intelligent data screening and caching.

[0042] In addition, it effectively guarantees the security and privacy of data and adopts advanced encryption and desensitizing technologies to provide all-round protection for user data.

[0043] Finally, through precise parameter value range selection methods, the accuracy of the algorithm and the performance of the model are greatly improved, providing more reliable support for various applications.

[0044] In summary, the artificial intelligence-based data interaction method and system of the present invention have significant advantages in data processing, user experience, system performance and data security, and have broad application prospects and market value. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a schematic diagram of the connection relationship of the system modules of an embodiment of the present application; DETAILED DESCRIPTION

[0046] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.

[0047] 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 the process, method, article or equipment 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 equipment. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the existence of other identical elements in the process, method, article or equipment including the elements.

[0048] In the embodiments of the present application, the same reference numerals represent the same components, and for the sake of brevity, detailed descriptions of the same components are omitted in different embodiments. It should be understood that the thickness, length, width and other dimensions of various components in the embodiments of the present application shown in the drawings, as well as the overall thickness, length, width and other dimensions of the integrated device are only exemplary descriptions and should not constitute any limitation to the present application.

[0049] See also Figure 1 As shown, Figure 1 Schematic diagram of the system module connection relationship of the embodiment of the present application.

[0050] The object of the present invention is to provide a data interaction method and system based on artificial intelligence, comprising:

[0051] The data acquisition module is used to collect raw data from various data sources such as databases, file systems, sensors, etc., and pass the collected data to the data preprocessing module;

[0052] The data preprocessing module is connected to the data acquisition module and is responsible for cleaning the received raw data, removing noise, duplication and erroneous data, converting data, unifying data in different formats into a standard format, and performing data normalization operations to make the data consistent and comparable. The processed data is transmitted to the artificial intelligence model module;

[0053] The artificial intelligence model module is connected to the data preprocessing module;

[0054] The data interaction module is connected to the artificial intelligence model module, and uses natural language processing technology to understand the natural language needs of users and convert them into executable query and operation instructions. It adopts a distributed computing framework to distribute data processing tasks to multiple computing nodes for parallel execution, dynamically adjusts the allocation of computing resources through load balancing algorithms, and uses intelligent data screening algorithms to predict data that users may be interested in based on user needs and historical behaviors, and sends relevant instructions to the cache module in advance for caching;

[0055] A cache module, connected to the data interaction module, is used to store the predicted data that the user may be interested in, so as to improve the response speed of data interaction;

[0056] The user interface module is connected to the data interaction module to provide users with a friendly operation interface, making it convenient for users to input requirements and obtain results;

[0057] A data management optimization module is connected to the data preprocessing module and the data interaction module respectively;

[0058] Intelligent screening module, connected with the data interaction module;

[0059] A parameter selection module connected to the artificial intelligence model module;

[0060] The security and permission management module is connected to the above modules and adopts encryption technology and data desensitization methods to ensure the security and privacy of data during transmission and storage. At the same time, it supports users to flexibly set data access rights to ensure user data security.

[0061] Specifically, the AI ​​model module can use the back-propagation algorithm to calculate the gradient L according to Formula 1 based on data features and user behavior, and adjust the weight of the neural network based on the gradient L. The parameter adjustment method is based on the gradient descent method, and the step size of the parameter update is controlled by setting an appropriate learning rate;

[0062]

[0063] Where N is the number of samples, y i is the true value, is the predicted value.

[0064] Furthermore, the data management optimization module uses data classification strategies to classify data, adjusts parameters by changing the initial value of the cluster center and the number of iterations, and optimizes the clustering effect. The classification formula is implemented using Formula 2:

[0065]

[0066] Among them, x i and x j are two data points and n is the number of features.

[0067] Specifically, the intelligent screening module uses a collaborative filtering algorithm to predict the data that users are interested in based on the user's historical behavior and the preferences of similar users, and optimizes the parameters by adjusting the weights or thresholds of the commonly evaluated items. The similarity calculation formula is shown in Formula 3:

[0068]

[0069] Among them, u and v are two users, I uv is the set of items that users u and v have evaluated together, r ui and r vi are the user and the rating of the item respectively.

[0070] Specifically, the data interaction module can dynamically optimize distributed computing and load balancing strategies based on real-time data traffic and the load of computing nodes.

[0071] Specifically, in the data interaction module, real-time monitoring of data flow and computing node load is the basis for dynamic optimization. By deploying monitoring tools and sensors in the system, relevant indicators of data flow, such as data input, output, and processing volume per second, as well as computing node load parameters, including CPU utilization, memory usage, disk I / O waiting time, etc., are continuously collected.

[0072] Specifically, when a new data processing task is received, the data interaction module first analyzes the current load status of each computing node. If the load of some nodes is close to or has reached saturation, while other nodes still have available resources, the new task will be assigned to the node with a lighter load for execution. This allocation decision not only takes into account the current load situation, but also estimates the subsequent data traffic trend based on the prediction model.

[0073] Specifically, for distributed computing tasks, the task segmentation and allocation strategy are dynamically adjusted according to the characteristics of the data and computing requirements. For example, if the data is highly correlated, the task segmentation will be reduced to reduce the data transmission overhead between nodes; if the data is relatively independent, the task can be segmented more finely and processed in parallel to improve efficiency.

[0074] Specifically, in terms of load balancing, a dynamic weight adjustment method is adopted. Different weights are assigned to each computing node based on the node's performance indicators (such as the number of CPU cores, memory size, etc.) and real-time load conditions. When assigning tasks, tasks are assigned to each node according to the weight ratio to ensure full utilization of resources and balanced distribution of load.

[0075] At the same time, the data interaction module will also monitor the execution of tasks in real time. If it is found that the load of a certain node increases or decreases abnormally during the execution of a task, the allocation of subsequent tasks will be adjusted in time, and even some tasks being executed will be migrated to other nodes with lighter loads to ensure the stable operation and efficient processing of the entire system.

[0076] Through the above detailed strategies and mechanisms, the data interaction module can achieve dynamic optimization according to the real-time data flow and computing node load, improve the effects of distributed computing and load balancing, and ensure the high performance and reliability of the system. Those skilled in the art can design and implement the corresponding system in accordance with the above description.

[0077] Specifically, the data management optimization module can automatically adjust data classification, compression and archiving strategies according to data growth and change trends.

[0078] Specifically, the data management optimization module will continuously monitor the growth and change trend of data. First, for data growth, it is evaluated by regularly analyzing indicators such as the speed of increase in data volume, the frequency of new data generation, and the occupancy of storage space.

[0079] Specifically, in terms of data classification, as data grows and changes, the module will reanalyze the characteristics and attributes of the data. For example, if the newly generated data has specific patterns or characteristics that are different from the original classification, the module will automatically create new classification categories or adjust the boundaries and rules of existing classifications. For originally clear classifications, if the characteristics of the data gradually become blurred or the overlap increases, the module may merge some similar classifications to simplify management.

[0080] Specifically, in terms of data compression, a more suitable compression algorithm is selected or compression parameters are adjusted according to the type of data and the change of repetitive patterns. If the new data contains more repeated information, an algorithm with a higher compression ratio will be used; and for data whose data patterns become complex and difficult to compress effectively, the compression strength may be reduced to avoid excessive consumption of computing resources, while ensuring a certain compression effect to save storage space.

[0081] Specifically, in terms of data archiving, adjustments are made based on the frequency of use and the changing trends of the importance of the data. For data whose frequency of use has significantly decreased but still has a certain value, timely archiving operations are performed and it is transferred to a storage medium with lower costs. At the same time, for data that was originally archived but has become important and frequently used again due to business needs, it is automatically restored to a high-speed storage area.

[0082] In order to implement these automatic adjustment strategies, the module will use the prediction model in machine learning to predict future trends based on the growth and change patterns of historical data. At the same time, combined with the real-time monitoring and analysis results of the data, the adjustment strategies and algorithms are continuously optimized to ensure the efficiency and adaptability of data management. Those skilled in the art can carry out corresponding development and configuration in the actual data management system based on the above description.

[0083] Specifically, the intelligent screening module can combine real-time user behavior and new user preference data to dynamically update the user interest prediction model.

[0084] Specifically, first, real-time monitoring of user behavior, including but not limited to user browsing history, click operations, search keywords, purchase behavior, collection actions, and dwell time in the app, etc. These real-time behavior data are continuously collected and transmitted to the intelligent screening module.

[0085] Specifically, at the same time, new user preference data is continuously acquired, which may come from the user's actively updated personal profile, feedback on recommended content (such as likes, dislikes, comments, etc.), and preference information reflected in social interactions with other users.

[0086] Specifically, for the collected real-time user behavior and new user preference data, the intelligent screening module uses data preprocessing technology to clean and normalize them to remove noise and abnormal data.

[0087] Specifically, in terms of model updates, online learning or incremental learning algorithms are used. For example, a gradient-based online learning method is used to calculate the update amount of model parameters in real time based on new data. For collaborative filtering-based models, when new user behaviors are generated, the similarity matrix between users is quickly recalculated and the recommendation list is updated.

[0088] Specifically, for deep learning models, such as models that use convolutional neural networks (CNN) or recurrent neural networks (RNN) to capture the sequential characteristics of user behavior, the weights of the model are fine-tuned according to new data through the back-propagation algorithm.

[0089] Specifically, in order to ensure the effectiveness and accuracy of the update, model evaluation indicators such as accuracy, recall, F1 value, etc. will be introduced to continuously monitor model performance during the update process. If the update causes the model performance to deteriorate, measures such as rollback or adjustment of the update strategy will be taken.

[0090] Specifically, through the above detailed steps and technical means, the intelligent screening module can timely and accurately combine real-time user behavior and new user preference data, dynamically update the user interest prediction model, and provide a clear reference execution path for technical personnel in this field.

[0091] Specifically, the parameter selection module can evaluate the impact of parameter selection on model performance in real time during model training, and dynamically adjust the parameter value range. For example, based on experience, the initial value range of the learning rate is set to [0.001, 0.1], and then the optimal value is further screened within this range through cross-validation.

[0092] The parameter selection module plays a key role in model training. The following is its detailed workflow and method:

[0093] Specifically, at the beginning of model training, the initial range of key parameters is set based on past experience and domain knowledge. Taking the learning rate as an example, its initial range is set to [0.001, 0.1] based on general practical experience.

[0094] Specifically, during the training process, the impact of parameter selection on model performance in each training batch or iteration is evaluated in real time. Evaluation indicators can include but are not limited to accuracy, recall, mean square error, etc., depending on the model's tasks and application scenarios.

[0095] Specifically, in order to more accurately determine the optimal parameter value, a cross-validation technique is used to divide the data set into multiple subsets, for example, in K-fold cross-validation, the data is divided into K subsets.

[0096] Specifically, in the set initial value range [0.001, 0.1], different learning rate values ​​are selected for training and validation. For each learning rate value, in K cross-validations, K-1 subsets are used for training in turn, and the remaining 1 subset is used for validation, and the average value of the evaluation index on the validation set is calculated.

[0097] Specifically, compare the average values ​​of the validation set evaluation indicators corresponding to different learning rate values, and select the learning rate value with the best performance as the current optimal value.

[0098] Specifically, if the current optimal value still does not meet the expected performance standard, or the model performance fluctuates in subsequent training, the parameter selection module will dynamically adjust the parameter range. For example, if the current optimal learning rate is 0.01 but the performance still needs to be improved, the range may be expanded to [0.0001, 0.5], and the above cross-validation and screening process will be performed again.

[0099] Specifically, throughout the entire process, the parameter selection module continuously monitors the model performance and continuously optimizes the parameter value range based on real-time feedback until it finds the parameter value combination that best suits the current data and model architecture to maximize the model performance.

[0100] Specifically, those skilled in the art can apply and implement the parameter selection module in actual model training according to the above steps and methods, thereby effectively improving the performance and generalization ability of the model.

[0101] Specifically, the user interface module can adaptively adjust the interface layout and function display according to the user's usage habits and operating preferences.

[0102] Specifically, firstly, the user's various operation behaviors and interaction data are continuously recorded during the user's use of the system. Such data include but are not limited to the function buttons that the user frequently clicks, the page areas that the user frequently visits, the time and frequency of operations, the order in which specific functions are used, and the location and duration of stay on the interface, etc.;

[0103] The collected user operation data is processed and mined using data analysis algorithms and machine learning techniques. For example, cluster analysis is used to group users with similar operation habits together, or association rule mining is used to discover potential associations between user operations.

[0104] In terms of interface layout adjustment, the position and size of functional modules are rearranged according to the user's operation frequency and importance;

[0105] Frequently used function buttons will be placed in more prominent and easy-to-operate locations, and their display size may be increased, while less frequently used functions may be reduced in size or hidden in secondary menus;

[0106] For function display, the display mode and priority of functions are dynamically adjusted according to the user's operation preferences and usage habits. For example, if a user often uses a specific filter condition or sorting method, the filter condition and sorting method will be used by default the next time the user enters the relevant page.

[0107] The interface module also takes into account the type of device the user uses (such as mobile phones, tablets, computers) and screen size to ensure that the adaptively adjusted interface can provide a good user experience on different devices.

[0108] In order to achieve a smooth transition and avoid giving users an abrupt feeling, the adjustment process can be gradual, gradually optimizing the interface layout and function display. At the same time, a user feedback mechanism is provided to allow users to manually adjust or restore default settings to meet personalized needs.

[0109] Through the above detailed technical steps and strategies, the user interface module can effectively make adaptive adjustments based on the user's usage habits and operating preferences, providing technical personnel with clear implementation ideas and execution methods.

[0110] Specifically, data transmission and interaction between modules follow efficient, stable and secure communication protocols.

[0111] In order to ensure that the data transmission and interaction between modules meet the requirements of efficiency, stability and security, the following communication protocol specifications are followed:

[0112] Efficiency:

[0113] Use data compression technology to compress data before transmission to reduce the amount of data and improve transmission efficiency. For example, use GZIP or Deflate compression algorithms.

[0114] The use of batch transmission and pipelining techniques allows multiple data requests or packets to be sent at a time, reducing round-trip delays and overhead.

[0115] Optimize data formats and choose lightweight and efficient data representation formats, such as JSON, Protobuf, etc., to reduce the time for data parsing and serialization.

[0116] Stability:

[0117] A heartbeat mechanism is established, and each module periodically sends heartbeat packets to detect the validity of the connection and promptly detect and handle connection interruptions or anomalies.

[0118] A retransmission and error correction mechanism is introduced to automatically retransmit and correct lost or damaged data to ensure data integrity.

[0119] Implement flow control and congestion control, dynamically adjust the data sending rate according to network conditions, and avoid data loss and increased delay caused by network congestion.

[0120] Security:

[0121] Use encrypted transmission, such as SSL / TLS protocol to encrypt transmitted data to prevent data from being stolen or tampered with.

[0122] Authentication and authorization are performed. Only authenticated and authorized modules can interact with data, ensuring the legitimacy of data access.

[0123] Implement data integrity checks, such as using a message digest algorithm (such as MD5, SHA-256) to check the transmitted data to verify that the data has not been modified during transmission.

[0124] In actual implementation, technicians can select and configure appropriate communication protocol stacks and related technical components based on specific system needs and performance requirements, and develop and optimize them according to the above principles to ensure that data transmission and interaction between modules are efficient, stable and secure.

[0125] Example 1: Personalized recommendation on e-commerce platform

[0126] 1. Data collection and preprocessing

[0127] A large amount of data is collected from multiple data sources such as user registration information, browsing history, purchase history, search keywords, etc. For example, user A's purchase history includes multiple purchases of sportswear, and his browsing history includes multiple searches for sports shoes.

[0128] Clean the collected data to remove duplicate and erroneous data. For example, remove duplicate purchase records and correct incorrect product classifications.

[0129] Use feature engineering technology to extract user behavior characteristics. For user A, the extracted features may include "sports category preference" and "high purchase frequency".

[0130] 2. Model training and optimization

[0131] Use collaborative filtering algorithm to build personalized recommendation model. Calculate the similarity between users. For example, the similarity between user A and user B is sim=(A,B). According to the formula Among them I AB is the set of products that users A and B have reviewed together. and are the ratings or purchase behaviors of user A and B on product i. If user B buys a new sports watch and gives it a high rating, and user A and B have a high similarity, the system will think that user A may also be interested in this sports watch.

[0132] At the same time, deep learning algorithms such as convolutional neural networks (CNN) are used to extract features from product images to further improve the accuracy of recommendations.

[0133] Training is performed based on users’ historical behavior data, and the model parameters are continuously adjusted to improve the accuracy of recommendations.

[0134] 3. Intelligent interaction and recommendation

[0135] When user A logs into the platform, the system uses natural language processing technology to understand the user's input "I want to buy summer sports equipment."

[0136] Combining user A's historical behavior and real-time input, the system uses the trained model to recommend relevant sports shorts, sports vests and other products to the user, and displays them on the user interface in the form of pictures and texts.

[0137] At the same time, the recommendation results are adjusted in real time based on users' real-time feedback, such as clicks, browsing time, etc.

[0138] 4. Effect evaluation

[0139] The effectiveness of the system can be evaluated by comparing indicators such as user click-through rate, purchase conversion rate, and average order value before and after the new system is adopted. If after adopting the new system, user A's click-through rate increases from 5% to 8%, and the purchase conversion rate increases from 2% to 3%, it means that the system is effective.

[0140] Collect user feedback, such as "The recommended products do not quite meet my needs", to further optimize the recommendation model and interaction methods.

Claims

1. A data interaction system based on artificial intelligence, It is characterized in that include: The data acquisition module is used to collect raw data from various data sources such as databases, file systems, sensors, etc., and pass the collected data to the data preprocessing module; The data preprocessing module is connected to the data acquisition module and is responsible for cleaning the received raw data, removing noise, duplication and erroneous data, converting data, unifying data in different formats into a standard format, and performing data normalization operations to make the data consistent and comparable. The processed data is transmitted to the artificial intelligence model module; The artificial intelligence model module is connected to the data preprocessing module; The data interaction module is connected to the artificial intelligence model module, and uses natural language processing technology to understand the natural language requirements input by users and convert them into executable query and operation instructions. It adopts a distributed computing framework to distribute data processing tasks to multiple computing nodes for parallel execution, dynamically adjusts the allocation of computing resources through load balancing algorithms, and uses intelligent data screening algorithms to predict data that users may be interested in based on user needs and historical behaviors, and sends relevant instructions to the cache module in advance for caching; A cache module, connected to the data interaction module, is used to store the predicted data that the user may be interested in, so as to improve the response speed of data interaction; The user interface module is connected to the data interaction module to provide users with a friendly operation interface, making it convenient for users to input requirements and obtain results; A data management optimization module is connected to the data preprocessing module and the data interaction module respectively; Intelligent screening module, connected with the data interaction module; A parameter selection module connected to the artificial intelligence model module; The security and permission management module is connected to the above modules and adopts encryption technology and data desensitization methods to ensure the security and privacy of data during transmission and storage. At the same time, it supports users to flexibly set data access rights to ensure user data security.

2. The artificial intelligence-based data interaction system according to claim 1, It is characterized in that The artificial intelligence model module can calculate the gradient L according to formula 1 using the back propagation algorithm based on data features and user behavior, and adjust the weight of the neural network according to the gradient L. The parameter adjustment method is based on the gradient descent method, and the step size of the parameter update is controlled by setting an appropriate learning rate; Where N is the number of samples, y i is the true value, is the predicted value.

3. The artificial intelligence-based data interaction system according to claim 1, It is characterized in that The data management optimization module uses data classification strategy to classify data, adjusts parameters by changing the initial value of the cluster center and the number of iterations, and optimizes the clustering effect. The classification formula is implemented using Formula 2: Among them, x i and x j are two data points and n is the number of features.

4. The artificial intelligence-based data interaction system according to claim 1, It is characterized in that The intelligent screening module uses a collaborative filtering algorithm to predict the data that the user is interested in based on the user's historical behavior and the preferences of similar users, and optimizes the parameters by adjusting the weights or thresholds of the commonly evaluated items. The similarity calculation formula is shown in Formula 3: Among them, u and v are two users, I uv is the set of items that users u and v have evaluated together, r ui and r vi are the user and the rating of the item respectively.

5. The data interaction system according to claim 1, It is characterized in that The data interaction module can dynamically optimize distributed computing and load balancing strategies according to real-time data traffic and the load conditions of computing nodes.

6. The data interaction system according to claim 3, It is characterized in that The data management optimization module can automatically adjust the strategies of data classification, compression and archiving according to the growth and change trend of data.

7. The data interaction system according to claim 4, It is characterized in that The intelligent screening module can combine real-time user behavior and new user preference data to dynamically update the user interest prediction model.

8. The data interaction system according to claim 1, It is characterized in that The parameter selection module can evaluate the impact of parameter selection on model performance in real time during model training, and dynamically adjust the parameter value range. For example, based on experience, the initial value range of the learning rate is set to [0.001, 0.1], and then the optimal value is further screened within this range through cross-validation.

9. The data interaction system according to claim 1, It is characterized in that The user interface module can adaptively adjust the interface layout and function display according to the user's usage habits and operating preferences.

Citation Information

Patent Citations

  • Low-conductivity working solution for optimizing neural network based on genetic algorithm and design method

    CN117252084A

  • Interactive artificial intelligence application platform based on big data

    CN117992577A

  • Intelligent personalized recommendation system

    CN118013107A

  • Commodity recommendation system based on artificial intelligence

    CN118134591A

  • Predictive asset optimization for computer resources

    JP2021005387A

Cited By

  • Data acquisition and processing system for user text emotional tendency analysis

    CN119721007A

  • Data Acquisition and Processing System for User Text Sentiment Tendency Analysis

    CN119721007B