Sharing platform intelligent recommendation method based on big data

Through intelligent recommendation methods and encryption technology based on big data, the problem of insufficient data security in the intelligent recommendation system of the shared platform is solved, and high security of data in transmission and storage is achieved, preventing tampering and theft, and reducing data leakage risk.

CN119066261APending Publication Date: 2024-12-03ANHUI UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410918943.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-10
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing shared platform intelligent recommendation system cannot effectively protect the security of data, especially financial-related data is easily tampered with and stolen, resulting in high risks in data transmission and storage.

Method used

Adopt intelligent recommendation methods based on big data, including data collection, cleaning and analysis, model construction and evaluation, encryption management and other steps, and ensure the security of data during storage and transmission through intelligent recommendation system tools and encryption technology, and use transparent data Encrypted, efficient data retrieval and highly disastrous technologies prevent data tampering and theft.

Benefits of technology

Effectively prevent malicious attackers from stealing or tampering with data, improve data security, reduce the risk of data leakage, ensure that data is not accessed unauthorized during transmission and storage, and protect the security of sensitive information.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119066261A_ABST
    Figure CN119066261A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of sharing platform intelligent recommendation, and discloses a sharing platform intelligent recommendation method based on big data, which comprises the following steps: S1, collecting and analyzing user data: providing information and insight required by decision making, and discovering rules and trends hidden behind the data by analyzing a large amount of data; s2, selecting a proper intelligent recommendation system tool: filtering out information which a user is not interested in according to the demand and preference of the user, and reducing information overload; the method has the beneficial effects that by setting encryption management and encryption technologies, malicious attackers are effectively prevented from stealing, tampering or damaging the data, the data security is improved, the data is protected from being influenced by the malicious attacks, the data is effectively prevented from being accessed by unauthorized people when being stolen or lost, the data leakage risk is reduced, the data security is protected, and the security of the data is improved. Unauthorized visitors are prevented from obtaining sensitive information, and it is ensured that data are not stolen or tampered in the transmission and storage process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent recommendation of sharing platforms, and in particular to an intelligent recommendation method of sharing platforms based on big data. Background Art

[0002] The significance of intelligent recommendation on sharing platform is to improve resource utilization efficiency, enhance service quality, promote fairness and inclusiveness, enhance social interaction and participation, and lead new consumption trends through personalized recommendations. Intelligent recommendation system can integrate social resources, improve resource utilization efficiency and public service level. However, the above scheme cannot enforce encryption of relevant data, and the data can be easily captured by illegal problems and exposed at a glance, especially passwords. Data related to finance is particularly important and easy to be tampered with. The content of the receiver and the sender is inconsistent. Summary of the invention

[0003] The purpose of the present invention is to provide a sharing platform intelligent recommendation method based on big data in order to solve the above problems, thereby solving the problems mentioned in the background technology.

[0004] In order to solve the above problems, the present invention provides a technical solution:

[0005] A sharing platform intelligent recommendation method based on big data, comprising the following steps:

[0006] S1: Collect and analyze user data: It can provide the information and insights needed for decision making. By analyzing big data, the patterns and trends hidden behind the data can be discovered;

[0007] S2: Choose the appropriate intelligent recommendation system tool: it can filter out information that users are not interested in according to their needs and preferences, thus reducing information overload;

[0008] S3: Training and optimizing intelligent recommendation models: Pre-training models on large amounts of data, usually including word embedding, convolutional neural networks, and recurrent neural networks;

[0009] S4: Recommendation engine: The recommendation engine is the core algorithm used to generate the recommendation list. The recommendation engine can be a content-based recommendation algorithm or a collaborative filtering-based recommendation algorithm:

[0010] S5: Encryption Management: Ensure the security of enterprise data during storage and transmission by protecting, sharing, backing up and organizing encryption keys;

[0011] S6: System function: responsible for controlling the creation, termination, scheduling and communication of processes to ensure the stable operation of the system.

[0012] As a preferred solution of the present invention, the collecting and analyzing user data in S1 mainly includes the following steps:

[0013] S11: Data collection: can come from various sources, such as internal databases, third-party data providers, social media platforms, etc. Data collection needs to consider data quality, scope and relevance;

[0014] S12: Data cleaning: The collected data often contains noise, missing values, outliers and duplicate data. The goal of data cleaning is to solve these problems to improve data quality and analysis accuracy;

[0015] S13: Data exploration: Visualization tools, statistical methods, etc. can be used to understand the distribution, relationships, and patterns of data;

[0016] S14: Model building: In the model building phase, we select appropriate algorithms and models based on the characteristics of the data and the analysis objectives;

[0017] S15: Model evaluation: Use various evaluation indicators, such as accuracy, recall, F1 score, etc., to measure the performance of the model. In addition, we can also optimize the performance of the model through cross-validation, adjusting hyperparameters, etc.

[0018] As a preferred solution of the present invention, the selecting of a suitable intelligent recommendation system tool according to S2 mainly comprises the following steps:

[0019] S21: Feature extraction and representation: When building an intelligent recommendation system, users and items need to be represented in a suitable vector form in order to calculate the similarity or correlation between them;

[0020] S22: Establishing a recommendation model: Establishing a recommendation model is one of the core tasks in the design of an intelligent recommendation system. The recommendation model uses the feature vectors of users and items to learn the association rules between them, and predicts the items that users may be interested in based on the learned knowledge;

[0021] S23: Evaluation and optimization: It is necessary to evaluate and optimize the system performance. Evaluation indicators can include accuracy, recall, coverage, etc. when building an intelligent recommendation system to measure the quality of recommendation results.

[0022] S24: Deployment and Application: After completing the construction and optimization of the intelligent recommendation system, the system needs to be deployed to actual application scenarios.

[0023] As a preferred solution of the present invention, the training and optimization of the intelligent recommendation model in S3 mainly includes the following steps:

[0024] S31: Model training: Model training is the process of allowing the machine to continuously optimize model parameters through data. The core of this step is the optimization algorithm, such as the gradient descent algorithm, the stochastic gradient descent algorithm, etc.

[0025] S32: Model evaluation: During or after training, we need to evaluate the model to understand its performance and accuracy;

[0026] S33: Adjust the model: Based on the evaluation results, we may need to adjust the model to improve its performance, which may include changing the model structure, modifying hyperparameters, adding or reducing features, etc. During the adjustment process, we need to constantly try and optimize to find the best model configuration;

[0027] S34: Model saving and deployment: After training and adjustment, we save the trained model for subsequent use. For easy use and sharing, we usually convert the model into a unified format and store it in cloud storage or local file system.

[0028] As a preferred solution of the present invention, the recommendation engine in S4 mainly includes the following steps:

[0029] S41: Personalized video recommendations: Video streaming platforms can use deep learning models, such as convolutional neural networks and recurrent neural networks, to make personalized video recommendations based on users’ viewing history and video content features;

[0030] S42: Real-time analysis: Analyze current popular videos and user viewing history to update recommended content in real time;

[0031] S43: Recommendation list generation: Recommendation list generation is the final task in the recommendation system, which involves generating a recommendation list containing a certain number of items.

[0032] As a preferred solution of the present invention, the encryption management in S5 mainly includes the following steps:

[0033] S51: Transparent data encryption: It is completely transparent to applications and tools, and does not require modification. It can provide three granularities for databases: columns, tables, and tablespaces.

[0034] S52: Efficient data retrieval: After the database encrypts the data, it can still provide indexing capabilities for the ciphertext data, thereby maintaining efficient access to the database;

[0035] S53: High availability and easy maintenance: The product uses technologies such as data integration storage with the database, RAC support, dual-machine hot standby, emergency mode, multi-process redundancy, transparent fault switching, data error ignoring, backup and recovery, etc.

[0036] S54: High disaster tolerance: It does not affect the database recovery, backup, synchronization and other operations of the database itself, and provides additional offline data recovery tools to ensure that the data can be restored to its original state even in extreme cases, thus ensuring high data availability.

[0037] As a preferred solution of the present invention, the system function in S6 mainly includes the following steps:

[0038] S61: Service resource management: Integrate various service resources, including human resources, material resources, technical resources, etc., to achieve unified management and deployment of resources;

[0039] S62: Service process management: Customize service processes, set different service stages and tasks according to enterprise needs, and realize automated processing of service processes, including task allocation, execution, monitoring and feedback;

[0040] S63: User management: supports multi-user role and permission management to meet the usage requirements of different departments and positions in the enterprise.

[0041] The beneficial effects of the present invention are: by setting up encryption management, encryption technology can effectively prevent malicious attackers from stealing, tampering with or destroying data, improve data security, protect data from malicious attacks, effectively prevent data from being accessed by unauthorized persons when it is stolen or lost, reduce the risk of data leakage, protect data security, prevent unauthorized visitors from obtaining sensitive information, and ensure that data is not stolen or tampered with during transmission and storage. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] For ease of explanation, the present invention is described in detail with reference to the following specific implementations and the accompanying drawings.

[0043] Figure 1 It is a flow chart of intelligent recommendation of the sharing platform of the present invention. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0045] Example 1

[0046] See also Figure 1 The present invention provides a technical solution: a sharing platform intelligent recommendation method based on big data, comprising the following steps:

[0047] S1: Collect and analyze user data: It can provide the information and insights needed for decision making. By analyzing large amounts of data, the patterns and trends hidden behind the data can be discovered.

[0048] S2: Choose the right intelligent recommendation system tool: It can filter out information that users are not interested in based on their needs and preferences, thus reducing information overload;

[0049] S3: Training and optimizing intelligent recommendation models: Models pre-trained on large amounts of data, usually including word embeddings, convolutional neural networks, and recurrent neural networks;

[0050] S4: Recommendation engine: The recommendation engine is the core algorithm used to generate the recommendation list. The recommendation engine can be a content-based recommendation algorithm or a collaborative filtering-based recommendation algorithm:

[0051] S5: Encryption Management: Ensure the security of enterprise data during storage and transmission by protecting, sharing, backing up and organizing encryption keys;

[0052] S6: System function: responsible for controlling the creation, termination, scheduling, and communication of processes to ensure the stable operation of the system.

[0053] Example 2

[0054] The collecting and analyzing of user data in S1 mainly includes the following steps:

[0055] S11: Data collection: can come from various sources, such as internal databases, third-party data providers, social media platforms, etc. Data collection needs to consider data quality, scope and relevance;

[0056] S12: Data cleaning: The collected data often contains noise, missing values, outliers and duplicate data. The goal of data cleaning is to solve these problems to improve data quality and analysis accuracy;

[0057] S13: Data exploration: Visualization tools, statistical methods, etc. can be used to understand the distribution, relationships, and patterns of data;

[0058] S14: Model building: In the model building phase, we select appropriate algorithms and models based on the characteristics of the data and the analysis objectives;

[0059] S15: Model evaluation: Use various evaluation indicators, such as accuracy, recall, F1 score, etc., to measure the performance of the model. In addition, we can also optimize the performance of the model through cross-validation, adjusting hyperparameters, etc.

[0060] An effective data source is crucial to the quality of analysis. Ensuring the accuracy, completeness, and representativeness of data is the core goal of the data collection phase. In addition, considering the possible bias in the data, adopting an appropriate sampling method is also a key consideration. The goal of collating data is to convert it into a form suitable for analysis to facilitate subsequent modeling and exploratory data analysis.

[0061] Example 3

[0062] The selection of a suitable intelligent recommendation system tool according to S2 mainly includes the following steps:

[0063] S21: Feature extraction and representation: When building an intelligent recommendation system, users and items need to be represented in a suitable vector form in order to calculate the similarity or correlation between them;

[0064] S22: Establishing a recommendation model: Establishing a recommendation model is one of the core tasks in the design of an intelligent recommendation system. The recommendation model uses the feature vectors of users and items to learn the association rules between them, and predicts the items that users may be interested in based on the learned knowledge;

[0065] S23: Evaluation and optimization: It is necessary to evaluate and optimize the system performance. Evaluation indicators can include accuracy, recall, coverage, etc. when building an intelligent recommendation system to measure the quality of recommendation results.

[0066] S24: Deployment and Application: After completing the construction and optimization of the intelligent recommendation system, the system needs to be deployed to actual application scenarios.

[0067] You can choose appropriate tools and technologies to build an efficient intelligent recommendation system and realize personalized recommendation services. This not only involves data processing and feature extraction, but also includes the establishment and optimization of recommendation models. It is a key link in the design and implementation of intelligent recommendation systems.

[0068] Example 4

[0069] The training and optimization of the intelligent recommendation model according to S3 mainly includes the following steps:

[0070] S31: Model training: Model training is the process of allowing the machine to continuously optimize model parameters through data. The core of this step is the optimization algorithm, such as the gradient descent algorithm, the stochastic gradient descent algorithm, etc.

[0071] S32: Model evaluation: During or after training, we need to evaluate the model to understand its performance and accuracy;

[0072] S33: Adjust the model: Based on the evaluation results, we may need to adjust the model to improve its performance, which may include changing the model structure, modifying hyperparameters, adding or reducing features, etc. During the adjustment process, we need to constantly try and optimize to find the best model configuration;

[0073] S34: Model saving and deployment: After training and adjustment, we save the trained model for subsequent use. For easy use and sharing, we usually convert the model into a unified format and store it in cloud storage or local file system.

[0074] Continuous adjustment and optimization are required. By following the steps of data preprocessing, model selection, model training, model evaluation and adjustment, we can train an AI model with excellent performance and play an important role in actual scenarios.

[0075] Example 5

[0076] The recommendation engine according to S4 mainly includes the following steps:

[0077] S41: Personalized video recommendations: Video streaming platforms can use deep learning models, such as convolutional neural networks and recurrent neural networks, to make personalized video recommendations based on users’ viewing history and video content features;

[0078] S42: Real-time analysis: Analyze current popular videos and user viewing history to update recommended content in real time;

[0079] S43: Recommendation list generation: Recommendation list generation is the final task in the recommendation system, which involves generating a recommendation list containing a certain number of items.

[0080] The recommendation engine can understand the content of each article more accurately, further improving the accuracy of the recommendation. This enables the recommendation engine to better match the user's interests with the characteristics of the article to generate a more personalized recommendation list. The personalized recall algorithm is the core of the recommendation engine, which pays more attention to the user's personalized needs and interests.

[0081] Example 6

[0082] The encryption management according to S5 mainly includes the following steps:

[0083] S51: Transparent data encryption: It is completely transparent to applications and tools, without the need for modification, and the database can be provided at the list granularity;

[0084] S52: Efficient data retrieval: After the database encrypts the data, it can still provide indexing capabilities for the ciphertext data, thereby maintaining efficient access to the database;

[0085] S53: High availability and easy maintenance: The product uses technologies such as data integration storage with the database, RAC support, dual-machine hot standby, emergency mode, multi-process redundancy, transparent fault switching, data error ignoring, backup and recovery, etc.

[0086] S54: High disaster tolerance: It does not affect the database recovery, backup, synchronization and other operations of the database itself, and provides additional offline data recovery tools to ensure that the data can be restored to its original state even in extreme cases, thus ensuring high data availability.

[0087] According to the security requirements and performance requirements of the data, select a suitable encryption algorithm. The key is a parameter that controls the encryption and decryption process. It needs to be long enough and random enough to improve the security of encryption. The encryption algorithm and key are used to convert plaintext data into ciphertext. Ciphertext is an encoded form. The ciphertext is securely stored or transmitted over the network to the destination. The receiver uses the correct key and decryption algorithm to convert the key into plaintext data.

[0088] Example 7

[0089] The system functions in S6 mainly include the following steps:

[0090] S61: Service resource management: Integrate various service resources, including human resources, material resources, technical resources, etc., to achieve unified management and deployment of resources;

[0091] S62: Service process management: Customize service processes, set different service stages and tasks according to enterprise needs, and realize automated processing of service processes, including task allocation, execution, monitoring and feedback;

[0092] S63: User management: supports multi-user role and permission management to meet the usage requirements of different departments and positions in the enterprise.

[0093] To sum up, the working principle of this technical solution is as follows: an effective data source is crucial to the quality of analysis. Ensuring the accuracy, completeness and representativeness of the data is the core goal of the data collection stage. In addition, considering the possible deviations in the data, adopting an appropriate sampling method is also a key consideration. The goal of collating the data is to convert it into a form suitable for analysis for subsequent modeling and exploratory data analysis. Appropriate tools and techniques can be selected to build an efficient intelligent recommendation system to achieve personalized recommendation services, which not only involves data processing and feature extraction, but also includes the establishment and optimization of recommendation models. It is a key link in the design and implementation of intelligent recommendation systems and requires continuous adjustment and optimization. By following steps such as data preprocessing, model selection, model training, model evaluation and adjustment, we can train an artificial intelligence model with excellent performance and play an important role in actual scenarios. The recommendation engine can understand the content of each article more accurately and further improve the accuracy of recommendations. This enables the recommendation engine to better match the user's interests with the characteristics of the article to generate a more personalized recommendation list. The personalized recall algorithm is the core of the recommendation engine. It pays more attention to the user's personalized needs and interests. According to the security requirements and performance requirements of the data, a suitable encryption algorithm is selected. The key is a parameter that controls the encryption and decryption process. It is necessary to ensure that it is long enough and random to improve the security of encryption. The encryption algorithm and key are used to convert plaintext data into ciphertext. Ciphertext is an encoded form. The ciphertext is securely stored or transmitted to the receiving location over the network. The recipient uses the correct key and decryption algorithm to convert the key into plaintext data.

[0094] 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. A sharing platform intelligent recommendation method based on big data, characterized in that: The following steps are involved: S1: Collect and analyze user data: It can provide the information and insights needed for decision making. By analyzing large amounts of data, the patterns and trends hidden behind the data can be discovered. S2: Choose the right intelligent recommendation system tool: It can filter out information that users are not interested in based on their needs and preferences, thus reducing information overload; S3: Training and optimizing intelligent recommendation models: Models pre-trained on large amounts of data, usually including word embeddings, convolutional neural networks, and recurrent neural networks; S4: Recommendation engine: The recommendation engine is the core algorithm used to generate the recommendation list. The recommendation engine can be a content-based recommendation algorithm or a collaborative filtering-based recommendation algorithm: S5: Encryption Management: Ensure the security of enterprise data during storage and transmission by protecting, sharing, backing up and organizing encryption keys; S6: System function: responsible for controlling the creation, termination, scheduling, and communication of processes to ensure the stable operation of the system.

2. The method for intelligent recommendation of a sharing platform based on big data according to claim 1, characterized in that: The collecting and analyzing of user data in S1 mainly includes the following steps: S11: Data collection: can come from various sources, such as internal databases, third-party data providers, social media platforms, etc. Data collection needs to consider data quality, scope and relevance; S12: Data cleaning: The collected data often contains noise, missing values, outliers and duplicate data. The goal of data cleaning is to solve these problems to improve data quality and analysis accuracy; S13: Data exploration: Visualization tools, statistical methods, etc. can be used to understand the distribution, relationships, and patterns of data; S14: Model building: In the model building phase, we select appropriate algorithms and models based on the characteristics of the data and the analysis objectives; S15: Model evaluation: Use various evaluation indicators, such as accuracy, recall, F1 score, etc., to measure the performance of the model. In addition, we can also optimize the performance of the model through cross-validation, adjusting hyperparameters, etc.

3. The intelligent recommendation method for a sharing platform based on big data according to claim 1, characterized in that: The selection of a suitable intelligent recommendation system tool according to S2 mainly includes the following steps: S21: Feature extraction and representation: When building an intelligent recommendation system, users and items need to be represented in a suitable vector form in order to calculate the similarity or correlation between them; S22: Establishing a recommendation model: Establishing a recommendation model is one of the core tasks in the design of an intelligent recommendation system. The recommendation model uses the feature vectors of users and items to learn the association rules between them, and predicts the items that users may be interested in based on the learned knowledge; S23: Evaluation and optimization: It is necessary to evaluate and optimize the system performance. Evaluation indicators can include accuracy, recall, coverage, etc. when building an intelligent recommendation system to measure the quality of recommendation results. S24: Deployment and Application: After completing the construction and optimization of the intelligent recommendation system, the system needs to be deployed to actual application scenarios.

4. The intelligent recommendation method for a sharing platform based on big data according to claim 1, characterized in that: The training and optimization of the intelligent recommendation model according to S3 mainly includes the following steps: S31: Model training: Model training is the process of allowing the machine to continuously optimize model parameters through data. The core of this step is the optimization algorithm, such as the gradient descent algorithm, the stochastic gradient descent algorithm, etc. S32: Model evaluation: During or after training, we need to evaluate the model to understand its performance and accuracy; S33: Adjust the model: Based on the evaluation results, we may need to adjust the model to improve its performance, which may include changing the model structure, modifying hyperparameters, adding or reducing features, etc. During the adjustment process, we need to constantly try and optimize to find the best model configuration; S34: Model saving and deployment: After training and adjustment, we save the trained model for subsequent use. For easy use and sharing, we usually convert the model into a unified format and store it in cloud storage or local file system.

5. The intelligent recommendation method for a sharing platform based on big data according to claim 1, characterized in that: The recommendation engine according to S4 mainly includes the following steps: S41: Personalized video recommendations: Video streaming platforms can use deep learning models, such as convolutional neural networks and recurrent neural networks, to make personalized video recommendations based on users’ viewing history and video content features; S42: Real-time analysis: Analyze current popular videos and user viewing history to update recommended content in real time; S43: Recommendation list generation: Recommendation list generation is the final task in the recommendation system, which involves generating a recommendation list containing a certain number of items.

6. The method for intelligent recommendation of a sharing platform based on big data according to claim 1, characterized in that: The encryption management according to S5 mainly includes the following steps: S51: Transparent data encryption: It is completely transparent to applications and tools, without the need for modification, and the database can be provided at the list granularity; S52: Efficient data retrieval: After the database encrypts the data, it can still provide indexing capabilities for the ciphertext data, thereby maintaining efficient access to the database; S53: High availability and easy maintenance: The product uses technologies such as data integration storage with the database, RAC support, dual-machine hot standby, emergency mode, multi-process redundancy, transparent fault switching, data error ignoring, backup and recovery, etc. S54: High disaster tolerance: It does not affect the database recovery, backup, synchronization and other operations of the database itself, and provides additional offline data recovery tools to ensure that the data can be restored to its original state even in extreme cases, thus ensuring high data availability.

7. The method for intelligent recommendation of a sharing platform based on big data according to claim 1, characterized in that: The system functions in S6 mainly include the following steps: S61: Service resource management: Integrate various service resources, including human resources, material resources, technical resources, etc., to achieve unified management and deployment of resources; S62: Service process management: Customize service processes, set different service stages and tasks according to enterprise needs, and realize automated processing of service processes, including task allocation, execution, monitoring and feedback; S63: User management: supports multi-user role and permission management to meet the usage requirements of different departments and positions in the enterprise.