Dynamic scheduling method based on fuel gas service middle station

Abstract

According to the dynamic scheduling method based on the fuel gas service platform, through data collection and preprocessing, the system can automatically collect, screen and sort various types of fuel gas service data, the time and energy of manual screening of a user are reduced, meanwhile, the system supports data of various data sources and forms, including texts, audios and videos, and the system is convenient to use. The system can quickly respond and meet the requirements of different users for gas use, and through the reinforcement learning analysis module and the knowledge graph module, the system can effectively filter out repeated or low-quality content, help the users to quickly find the most relevant and valuable gas package, and improve the user experience. The knowledge graph technology provides a concept association view between the data for a user, helps the user better understand the content and mutual relation of the data, improves the utilization efficiency of the gas, and meanwhile, the system has a personalized recommendation function.

Description

Technical Field

[0001] The present invention relates to the field of gas business technology, and in particular to a dynamic scheduling method based on a gas business middle station. Background Art

[0002] With the adjustment of the global energy structure and the advancement of my country's energy policy, the gas industry is facing an important opportunity for transformation and upgrading. As one of the important directions to promote the transformation and upgrading of the energy industry, smart gas construction has become an industry consensus.

[0003] Smart gas construction realizes intelligent transformation and management of gas systems by introducing advanced technologies such as the Internet of Things, big data, and cloud computing, improves the safety, stability, and efficiency of the system, and achieves rational use of resources and sustainable development of the environment. However, traditional gas systems have the following disadvantages:

[0004] 1. Reliance on manual work to handle gas business: The existing system mostly relies on manual input and sorting to handle gas business. This method is time-consuming and inefficient, takes a lot of time, and is difficult to cope with rapidly changing needs.

[0005] 2. Information overload and content duplication: Due to the diversity of gas services and the huge amount of information, manual adjustments to business dynamics often face the problems of information overload and duplication. The existing system fails to effectively filter and optimize recommended gas services, resulting in users having difficulty in quickly locating and obtaining truly valuable resources in the process of choosing gas services.

[0006] 3. Insufficient personalized recommendations and adaptability: The dynamic scheduling of the existing gas business platform lacks sufficient personalized recommendation functions and fails to customize the push of gas services according to the user's style, characteristics and needs, which makes it difficult for the system to meet the diverse needs of different users.

[0007] 4. Technology integration and user experience: Some existing systems have deficiencies in technology integration and user experience, such as unfriendly front-end interface design and limited back-end data processing capabilities, which affect the user experience and overall system performance.

[0008] Therefore, it is necessary to provide a dynamic scheduling method based on the gas business middle platform to solve the above technical problems. Summary of the invention

[0009] In order to solve the above technical problems, the present invention provides a dynamic scheduling method based on the gas business middle station to solve the need to accurately identify all personnel at the same time, resulting in the need to process large amounts of data in real time at the same time, the required amount of calculation is large and the real-time performance is not high, and when it comes to identification of different scenarios, it lacks applicability, and thus the efficiency of handling emergencies is low.

[0010] The present invention provides a dynamic scheduling method based on a gas service middle station, wherein the push method comprises the following steps:

[0011] S1. Collect and pre-process real-time data information of gas business;

[0012] S2. Build a knowledge graph based on data information to associate gas business data in real time;

[0013] S3, based on deep learning and feature selection technology, extract key features and train the gas business data recommendation model;

[0014] S4. Generate a gas service recommendation list using a recommendation algorithm and push strategy, and push it to the user;

[0015] S5. Implement a feedback mechanism to dynamically optimize the recommendation algorithm and push strategy based on user feedback.

[0016] The step S1 specifically includes the following steps:

[0017] S11. Gas business collection includes obtaining detailed information on gas business usage plans, usage targets, and usage arrangements, which are collected through gas business systems or online platforms;

[0018] S12. Use the Pandas library in Python to clean the data, including removing duplicates, processing missing data, and unifying the format to ensure the integrity and availability of the data;

[0019] S13. Perform word segmentation and part-of-speech tagging on the collected text data of gas business, extract keywords and topics using natural language processing technology, and identify content related to gas business needs;

[0020] S14, perform data normalization processing to unify data from different sources and formats into a certain standardized format to facilitate subsequent feature extraction and model training;

[0021] S15. Use data mining technology to analyze and mine historical gas business data to discover potential patterns and trends hidden in the data to support more accurate gas business recommendations;

[0022] S16. Based on the collected gas usage data, use machine learning or rule engine technology to establish a personalized gas demand model;

[0023] S17. Convert the gas usage demand description into a structured data representation so that the system can understand and process it;

[0024] S18. Use text mining and sentiment analysis techniques to analyze the text descriptions of users, identify their focus, preferences and special needs for gas business, and provide guidance for subsequent recommendation systems;

[0025] S19. Integrate gas data from different sources to ensure the integrity and consistency of the integrated data set;

[0026] S110. Implement data verification and quality control, including data legitimacy verification, logical consistency check and abnormal data processing to ensure data reliability and validity;

[0027] S111. Use visualization tools to conduct exploratory analysis on data to discover the distribution and potential associations of gas business data, providing a basis for subsequent knowledge graph construction and model training;

[0028] S112. Establish data management strategies and security measures to ensure the privacy protection and information security of users and user data.

[0029] Preferably, the step S2 specifically includes the following steps:

[0030] S21. Use natural language processing technology to semantically annotate and identify entities in gas business data, identify key concepts, topics and entities, and convert text information into structured knowledge graph data based on part-of-speech tagging and named entity recognition technology to facilitate subsequent relationship modeling and reasoning;

[0031] S22. Establish a conceptual hierarchical structure of the knowledge graph, including basic information of users and basic concepts of gas usage, and perform hierarchical representation;

[0032] S23. Use graph database to store and manage knowledge graph data, support complex relationship modeling and query operations, use reasoning algorithms to analyze implicit relationships and dependencies in knowledge graphs, and enhance the intelligence and recommendation accuracy of gas usage systems;

[0033] S24. Design an update mechanism for the knowledge graph, regularly update and maintain information on course content and gas business materials, and ensure the timeliness and completeness of the knowledge graph.

[0034] S25. Introduce an incremental update strategy to update the gas usage map based on new user requirements and usage needs to reflect the latest user usage data;

[0035] S26. Implement semantic reasoning function, analyze the logical relationship between the gas service needs of users and the relevant content in the knowledge graph through the reasoning engine, and provide support for the recommendation system;

[0036] S27. Develop an efficient search engine that supports complex semantic-based queries and information retrieval to improve the efficiency and accuracy of searching for gas business information.

[0037] Preferably, the step S3 specifically includes the following steps:

[0038] S31. Based on deep learning technology, convolutional neural network and recurrent neural network models are used for feature extraction;

[0039] S32. Design corresponding feature extractors according to different types of gas business data to identify and extract key features related to gas business needs;

[0040] S33. Select the appropriate deep learning model architecture and algorithm, and configure the model according to the characteristics of the data and the requirements of the task;

[0041] S34. Use TensorFlow or PyTorch framework to train and optimize the model, and adjust the model's hyperparameters and learning strategies to improve the model's generalization ability and recommendation accuracy.

[0042] S35. Introduce cross-validation technology and ensemble learning methods to integrate the prediction results of multiple models to improve the stability and reliability of gas business recommendations;

[0043] S36. Design evaluation indicators to evaluate the performance of the trained model;

[0044] S37. Use validation sets and test sets to verify the generalization ability of the model, and analyze the performance differences of the model in different scenarios. Optimize model parameters and recommendation algorithms based on feedback data and user evaluations, and continuously improve the intelligence level of the system and user experience.

[0045] S38. Implement the deployment and integration of the trained model and integrate it into the recommendation system to support real-time gas business recommendation services;

[0046] S39. Configure the service ports and interfaces of the model to achieve rapid model calling and data interaction, ensuring the efficiency and stability of the system;

[0047] S310. Design model update and upgrade strategies, monitor model operation status and performance indicators, and timely adjust and optimize system operation efficiency.

[0048] Preferably, the step S4 specifically includes the following steps:

[0049] S41. Implement TF-IDF, Word2Vec or Doc2Vec content-based recommendation algorithm to analyze the text content and semantic similarity of gas business information and recommend information that matches the needs of users;

[0050] S42. Combining keyword extraction and text clustering technology, discover hidden associations and information in gas business data to improve the accuracy and diversity of recommendations;

[0051] S43. Introduce user-based collaborative filtering algorithm to analyze the similarities and preferences between users and peers, and recommend information with similar gas business history;

[0052] S44. Implement project-based collaborative filtering algorithms to analyze the association and cross-use of gas business data and provide cross-disciplinary and cross-field recommendation services;

[0053] S45. Design a multi-objective optimization algorithm to balance the relevance and novelty of recommended information to meet the personalized gas service needs of users;

[0054] S46. Combine fuzzy logic or evolutionary algorithms to solve the trade-off between diversity and accuracy in the recommendation process and improve the adaptability and intelligence of the system.

[0055] Preferably, the step S7 specifically further includes the following push system implementation:

[0056] S471. User interface design: Design an intuitive and friendly user interface that supports personalized user settings and operations, enables browsing, searching and selecting gas business information, and uses React or Vuejs front-end frameworks to develop responsive design interfaces that adapt to different terminal devices and improve user experience and usability;

[0057] S472. Backend service development: Develop stable and efficient backend services to handle user requests and data interactions, support the invocation of recommendation algorithms and models, use the Flask or Django backend framework, implement the server-side API interface, and support dynamic loading of data and real-time updating of recommendation results;

[0058] S473, Data interface and external resource integration: Implement data interface with external gas resource library, obtain the latest gas business information and gas packages, expand the content coverage and information update speed of the system, design data synchronization and integration strategies, ensure the consistency and timeliness of data between the system and external resource library, and avoid information islands and outdated data problems;

[0059] S474. System performance optimization and security protection: Optimize system performance, including optimizing the loading speed of the front-end interface and improving the concurrent processing capabilities of back-end services, ensure the stability and response speed of the system, introduce data encryption and user identity authentication mechanisms, protect the security of users and their personal information, and comply with relevant data protection laws and regulations and privacy policy requirements.

[0060] Preferably, the step S7 specifically includes the following steps:

[0061] S51. Develop personalized recommendation strategies based on recommendation strategies, adjust recommendation results according to historical behaviors and real-time feedback of users, and provide dynamic adjustment and optimization of gas business information push services;

[0062] S52. Implement user feedback collection and analysis functions based on the feedback mechanism, collect users’ evaluations and opinions on recommended materials, and provide real-time data support for recommendation algorithms and system optimization.

[0063] The dynamic scheduling system based on the gas business middle station is characterized in that the push system includes:

[0064] Data collection and preprocessing module, used to collect real-time data information of gas business, including data cleaning, normalization and feature extraction to ensure data accuracy and consistency;

[0065] The knowledge graph construction module is used to build a knowledge graph using natural language processing technology to associate gas business data in real time. The knowledge graph includes course topics, knowledge points, user goals and related reference materials, and performs semantic annotation and relationship reasoning on them. The knowledge graph includes course topics, knowledge points, user goals and related reference materials, and performs semantic annotation and relationship reasoning on them.

[0066] Feature selection and model training module, based on deep learning and feature selection technology, extracts key features related to gas business needs, uses convolutional neural network and recursive neural network for model training, and establishes a gas business data recommendation model;

[0067] The recommendation algorithm and push strategy module implements a combination of content-based recommendation algorithm and collaborative filtering algorithm to generate a gas business recommendation list, using a multi-objective optimization algorithm to balance the relevance and novelty of the recommendation;

[0068] Push system real module, design and implement gas business data push system, including user interface, background service and data interface;

[0069] Feedback mechanism and system optimization module implement feedback mechanism, allowing users to evaluate and provide feedback on the pushed gas business information, dynamically adjust and optimize the recommendation algorithm and push strategy based on feedback data, improve the intelligence level of the system and user satisfaction, and regularly update and maintain the system to ensure its continued effectiveness and adapt to new user needs.

[0070] Preferably, the data collection and preprocessing module specifically includes:

[0071] Data collection module: responsible for collecting data from various data sources, real-time data information of gas business;

[0072] Data cleaning module, used to clean the collected data, remove noise, duplicate data, and handle missing values;

[0073] Data normalization module, standardizing data formats to ensure data consistency and comparability;

[0074] Feature extraction module, which extracts useful features from cleaned and normalized data for further analysis or model training;

[0075] The data management and storage module manages and stores processed data to ensure data security, accessibility and continuous updating.

[0076] Preferably, the knowledge graph construction module specifically includes:

[0077] Entity recognition and extraction module, used to identify and extract key entities in text and data information related to gas business;

[0078] The relation extraction and annotation module is used to extract the relations between entities from the text and annotate the relations;

[0079] The semantic annotation and reasoning module performs deep semantic annotation on entities and relationships in the knowledge graph and uses the reasoning engine to perform logical reasoning;

[0080] The graph update and maintenance module maintains the integrity and accuracy of the knowledge graph and updates and expands the graph content in a timely manner.

[0081] Visualization and application module: Provides visualization display and application interface of knowledge graph to facilitate users to understand and use graph information.

[0082] Preferably, the feature selection and model training module specifically includes:

[0083] Feature selection module, used to select the key features most relevant to model training from the extracted features, reduce redundancy and improve model performance;

[0084] Model building module for building deep learning models using convolutional neural networks and recurrent neural networks for data modeling;

[0085] Model training and verification module: train deep learning models and verify their performance to ensure the accuracy and generalization ability of the models;

[0086] Model evaluation and tuning module: evaluates the performance of the model and optimizes the model based on the evaluation results;

[0087] Model deployment and application module: deploy the trained model to the production environment and integrate it into the gas business recommendation system.

[0088] Compared with the related art, the dynamic scheduling method based on the gas business middle station provided by the present invention has the following beneficial effects:

[0089] 1. Through data collection and preprocessing, the system of the present invention can automatically collect, screen and organize various types of gas business data, reducing the time and energy of manual screening by users. At the same time, the system supports data from multiple sources and forms, including text, audio, and video, and can quickly respond to and meet the different user themes and rapidly changing usage needs;

[0090] 2. The present invention enhances the learning analysis module and the knowledge graph module, so that the system can effectively filter out duplicate or low-quality content, helping users to quickly find the most relevant and valuable gas packages. The knowledge graph technology provides users with a conceptual association view between materials, helping users to better understand the content and relationship of the materials, and improve the efficiency of resource utilization;

[0091] 3. The system of the present invention has a powerful personalized recommendation function, which can customize and push gas usage information according to the user's usage style, the user's gas usage characteristics and needs, thereby enhancing the user's pertinence and adaptability. The multi-language support module and localization function enable the system to adapt to the needs of users with different languages ​​and cultural backgrounds, thus expanding the scope of application of the system;

[0092] 4. The dynamic content generation module of the system of the present invention can update and generate new user content in real time to ensure that users always use the gas usage method. The personalized path planning module provides users with customized gas dynamic resource recommendations to assist users in differentiated use and improve user usage effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] Figure 1 This is a flow chart of the dynamic scheduling method based on the gas business middle station provided by the present invention. DETAILED DESCRIPTION

[0094] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0095] Embodiment 1

[0096] Based on the dynamic scheduling method of the gas business middle station, the push method includes the following steps:

[0097] S1. Collect and pre-process real-time data information of gas business;

[0098] S2. Build a knowledge graph based on data information to associate gas business data in real time;

[0099] S3, based on deep learning and feature selection technology, extract key features and train the gas business data recommendation model;

[0100] S4. Generate a gas service recommendation list using a recommendation algorithm and push strategy, and push it to the user;

[0101] S5. Implementing a feedback mechanism to dynamically optimize the recommendation algorithm and push strategy based on user feedback. The push method includes the following steps:

[0102] S1. Collect and pre-process real-time data information of gas business;

[0103] S2. Build a knowledge graph based on data information to associate gas business data in real time;

[0104] S3, based on deep learning and feature selection technology, extract key features and train the gas business data recommendation model;

[0105] S4. Generate a gas service recommendation list using a recommendation algorithm and push strategy, and push it to the user;

[0106] S5. Implement a feedback mechanism to dynamically optimize the recommendation algorithm and push strategy based on user feedback.

[0107] Specifically, the dynamic scheduling method based on the gas service middle station is characterized in that step S1 specifically includes the following steps:

[0108] S11. Gas business collection includes obtaining detailed information on gas business usage plans, usage targets, and usage arrangements, which are collected through gas business systems or online platforms;

[0109] S12. Use the Pandas library in Python to clean the data, including removing duplicates, processing missing data, and unifying the format to ensure the integrity and availability of the data;

[0110] S13. Perform word segmentation and part-of-speech tagging on the text data collected for gas business, extract keywords and topics using natural language processing technology, and identify content related to gas business needs

[0111] S14, perform data normalization processing to unify data from different sources and formats into a certain standardized format to facilitate subsequent feature extraction and model training;

[0112] S15. Use data mining technology to analyze and mine historical gas business data to discover potential patterns and trends hidden in the data to support more accurate gas business recommendations;

[0113] S16. Based on the collected gas usage data, use machine learning or rule engine technology to establish a personalized gas demand model;

[0114] S17. Convert the gas usage demand description into a structured data representation so that the system can understand and process it;

[0115] S18. Use text mining and sentiment analysis techniques to analyze the text descriptions of users, identify their focus, preferences and special needs for gas business, and provide guidance for subsequent recommendation systems;

[0116] S19. Integrate gas data from different sources to ensure the integrity and consistency of the integrated data set;

[0117] S110. Implement data verification and quality control, including data legitimacy verification, logical consistency check and abnormal data processing to ensure data reliability and validity;

[0118] S111. Use visualization tools to conduct exploratory analysis on data to discover the distribution and potential associations of gas business data, providing a basis for subsequent knowledge graph construction and model training;

[0119] S112. Establish data management strategies and security measures to ensure the privacy protection and information security of users and user data.

[0120] Step S2 specifically includes the following steps:

[0121] S21. Use natural language processing technology to semantically annotate and identify entities in gas business data, identify key concepts, topics and entities, and convert text information into structured knowledge graph data based on part-of-speech tagging and named entity recognition technology to facilitate subsequent relationship modeling and reasoning;

[0122] S22. Establish a conceptual hierarchical structure of the knowledge graph, including basic information of users and basic concepts of gas usage, and perform hierarchical representation;

[0123] S23. Use graph database to store and manage knowledge graph data, support complex relationship modeling and query operations, use reasoning algorithms to analyze implicit relationships and dependencies in knowledge graphs, and enhance the intelligence and recommendation accuracy of gas usage systems;

[0124] S24. Design an update mechanism for the knowledge graph, regularly update and maintain information on course content and gas business materials, and ensure the timeliness and completeness of the knowledge graph.

[0125] S25. Introduce an incremental update strategy to update the gas usage map based on new user requirements and usage needs to reflect the latest user usage data;

[0126] S26. Implement semantic reasoning function, analyze the logical relationship between the gas service needs of users and the relevant content in the knowledge graph through the reasoning engine, and provide support for the recommendation system;

[0127] S27. Develop an efficient search engine that supports complex semantic-based queries and information retrieval to improve the efficiency and accuracy of searching for gas business information.

[0128] Step S3 specifically includes the following steps:

[0129] S31. Based on deep learning technology, convolutional neural network and recurrent neural network models are used for feature extraction;

[0130] S32. Design corresponding feature extractors according to different types of gas business data to identify and extract key features related to gas business needs;

[0131] S33. Select the appropriate deep learning model architecture and algorithm, and configure the model according to the characteristics of the data and the requirements of the task;

[0132] S34. Use TensorFlow or PyTorch framework to train and optimize the model, and adjust the model's hyperparameters and learning strategies to improve the model's generalization ability and recommendation accuracy.

[0133] S35. Introduce cross-validation technology and ensemble learning methods to integrate the prediction results of multiple models to improve the stability and reliability of gas business recommendations;

[0134] S36. Design evaluation indicators to evaluate the performance of the trained model;

[0135] S37. Use validation sets and test sets to verify the generalization ability of the model, analyze the performance differences of the model in different scenarios, optimize model parameters and recommendation algorithms based on feedback data and user evaluation, and continuously improve the intelligence level of the system and user experience;

[0136] S38. Implement the deployment and integration of the trained model and integrate it into the recommendation system to support real-time gas business recommendation services;

[0137] S39. Configure the service ports and interfaces of the model to achieve rapid model calling and data interaction, ensuring the efficiency and stability of the system;

[0138] S310. Design model update and upgrade strategies, monitor model operation status and performance indicators, and timely adjust and optimize system operation efficiency.

[0139] Specifically, step S4 specifically includes the following steps:

[0140] S41. Implement TF-IDF, Word2Vec or Doc2Vec content-based recommendation algorithm to analyze the text content and semantic similarity of gas business information and recommend information that matches the needs of users;

[0141] S42. Combining keyword extraction and text clustering technology, discover hidden associations and information in gas business data to improve the accuracy and diversity of recommendations;

[0142] S43. Introduce user-based collaborative filtering algorithm to analyze the similarities and preferences between users and peers, and recommend information with similar gas business history;

[0143] S44. Implement project-based collaborative filtering algorithms to analyze the association and cross-use of gas business data and provide cross-disciplinary and cross-field recommendation services;

[0144] S45. Design a multi-objective optimization algorithm to balance the relevance and novelty of recommended information to meet the personalized gas service needs of users;

[0145] S46. Combine fuzzy logic or evolutionary algorithms to solve the trade-off between diversity and accuracy in the recommendation process and improve the adaptability and intelligence of the system.

[0146] Specifically, step S7 also includes the following push system implementation:

[0147] S471. User interface design: Design an intuitive and friendly user interface that supports user personalized settings and operations to enable browsing, searching and selecting of gas business information. Use the React or Vuejs front-end framework to develop a responsive interface that is compatible with different terminal devices and improves user experience and usability.

[0148] S472. Backend service development: Develop stable and efficient backend services to handle user requests and data interactions, support the invocation of recommendation algorithms and models, use the Flask or Django backend framework, implement the server-side API interface, and support dynamic loading of data and real-time updating of recommendation results;

[0149] S473, Data interface and external resource integration: Implement data interface with external gas resource library, obtain the latest gas business information and gas packages, expand the content coverage and information update speed of the system, design data synchronization and integration strategies, ensure the consistency and timeliness of data between the system and external resource library, and avoid information islands and outdated data problems;

[0150] By integrating with external resources through data interfaces, business processing can be carried out: involving business processing, appointment payment and information inquiry.

[0151] Business processing module: provides online application and progress inquiry functions for online registration, repair, management modification, transfer of ownership and other services.

[0152] Appointment and payment module: supports users to pay gas fees online through various payment methods such as Alipay, WeChat, bank cards, etc., and provides functions such as appointment payment and automatic deduction.

[0153] Information query module: provides query functions for user account information, payment records, gas usage and other information.

[0154] Online customer service module: provides real-time online customer service consultation and answer services through text, voice, video and other media.

[0155] S474, System performance optimization and security protection: Optimize system performance, including optimizing the loading speed of the front-end interface and improving the concurrent processing capacity of the back-end service, ensure the stability and response speed of the system, introduce data encryption and user identity authentication mechanisms, protect the security of users and their personal information, and comply with relevant data protection laws and regulations and privacy policy requirements;

[0156] Specifically, step S7 specifically includes the following steps:

[0157] S51. Develop personalized recommendation strategies based on recommendation strategies, adjust recommendation results according to historical behaviors and real-time feedback of users, and provide dynamic adjustment and optimization of gas business information push services;

[0158] S52. Implement user feedback collection and analysis functions based on the feedback mechanism, collect users’ evaluations and opinions on recommended materials, and provide real-time data support for recommendation algorithms and system optimization.

[0159] Specifically, in the development of personalized recommendation systems, first, through the development of refined recommendation strategies, the system can deeply analyze the historical behavior data of users, such as browsing habits, download records, and usage frequency, and combine real-time feedback information to flexibly adjust and optimize the recommendation results, and provide users with dynamically adjusted gas business information push services. At the same time, the system also introduces a comprehensive user feedback collection and analysis mechanism, which collects users' evaluations and opinions on recommended materials through multiple channels, and uses natural language processing and sentiment analysis technology to deeply analyze these feedback data. This not only provides real-time data support for the continuous optimization of the recommendation algorithm, but also lays the foundation for improving the overall functionality of the system and enhancing user experience, forming a closed-loop continuous improvement process to ensure that the system always adapts to changes in user needs and personalized needs.

[0160] The dynamic scheduling method based on the gas business middle platform includes:

[0161] Data collection and preprocessing module, used to collect real-time data information of gas business, including data cleaning, normalization and feature extraction to ensure data accuracy and consistency;

[0162] The knowledge graph construction module is used to build a knowledge graph using natural language processing technology to associate gas business data in real time. The knowledge graph includes course topics, knowledge points, user goals and related reference materials, and performs semantic annotation and relationship reasoning on them. The knowledge graph includes course topics, knowledge points, user goals and related reference materials, and performs semantic annotation and relationship reasoning on them.

[0163] Feature selection and model training module, based on deep learning and feature selection technology, extracts key features related to gas business needs, uses convolutional neural network and recursive neural network for model training, and establishes a gas business data recommendation model;

[0164] The recommendation algorithm and push strategy module implements a combination of content-based recommendation algorithm and collaborative filtering algorithm to generate a gas business recommendation list, using a multi-objective optimization algorithm to balance the relevance and novelty of the recommendation;

[0165] Push system real module, design and implement gas business data push system, including user interface, background service and data interface;

[0166] Feedback mechanism and system optimization module implement feedback mechanism, allowing users to evaluate and provide feedback on the pushed gas business information, dynamically adjust and optimize the recommendation algorithm and push strategy based on feedback data, improve the intelligence level of the system and user satisfaction, and regularly update and maintain the system to ensure its continued effectiveness and adapt to new user needs.

[0167] The data collection and preprocessing modules specifically include:

[0168] Data collection module: responsible for collecting data from various data sources, real-time data information of gas business;

[0169] Data cleaning module, used to clean the collected data, remove noise, duplicate data, and handle missing values;

[0170] Data normalization module, standardizing data formats to ensure data consistency and comparability;

[0171] Feature extraction module, which extracts useful features from cleaned and normalized data for further analysis or model training;

[0172] The data management and storage module manages and stores processed data to ensure data security, accessibility and continuous updating.

[0173] Specifically, the task of the data collection and preprocessing module is to obtain a diverse data set from multiple sources, including user usage and user background information. Subsequently, the module performs detailed preprocessing operations on the collected data, including but not limited to data cleaning to remove noise and erroneous information, normalization to ensure the dimensional consistency of the data, and advanced feature extraction to identify and extract key patterns and information in the data. These steps are designed to improve the accuracy, consistency and availability of the data, and ensure that a reliable data foundation is provided for subsequent intelligent analysis and decision-making. In addition, the module also standardizes and converts the data to be compatible with different data sources and system requirements, and implements data integrity checks to ensure that there are no missing data and logical errors.

[0174] The knowledge graph construction modules specifically include:

[0175] Entity recognition and extraction module, used to identify and extract key entities in text and data information related to gas business;

[0176] The relation extraction and annotation module is used to extract the relations between entities from the text and annotate the relations;

[0177] The semantic annotation and reasoning module performs deep semantic annotation on entities and relationships in the knowledge graph and uses the reasoning engine to perform logical reasoning;

[0178] The graph update and maintenance module maintains the integrity and accuracy of the knowledge graph and updates and expands the graph content in a timely manner.

[0179] Visualization and application module: Provides visualization display and application interface of knowledge graph to facilitate users to understand and use graph information.

[0180] Specifically, the knowledge graph construction module uses natural language processing technology to construct a knowledge graph, which links course content, user goals and gas business information to form a comprehensive graph containing course topics, knowledge points, user goals and related reference materials. At the same time, these contents are semantically annotated and relationally inferred to more clearly show their internal connections and logical structures.

[0181] The feature selection and model training modules specifically include:

[0182] Feature selection module, used to select the key features most relevant to model training from the extracted features, reduce redundancy and improve model performance;

[0183] Model building module for building deep learning models using convolutional neural networks and recurrent neural networks for data modeling;

[0184] Model training and verification module: train deep learning models and verify their performance to ensure the accuracy and generalization ability of the models;

[0185] Model evaluation and tuning module: evaluates the performance of the model and optimizes the model based on the evaluation results;

[0186] Model deployment and application module: deploy the trained model to the production environment and integrate it into the gas business recommendation system.

[0187] Specifically, the feature selection and model training module uses advanced deep learning technology and feature selection methods to extract key features closely related to gas business needs from massive data. This module applies convolutional neural networks (CNN) to process and identify complex patterns in the data, especially spatial features, and uses recurrent neural networks (RNN) to capture and analyze the time series information and dependencies of the data. Through the collaborative work of these neural networks, the system conducts model training, gradually optimizes and builds a high-precision gas business data recommendation model. This model can comprehensively consider the personalized needs of users and the user environment, and provide them with accurate and personalized gas package recommendations. The module also includes hyperparameter optimization and cross-validation steps to ensure the robustness and generalization ability of the model.

[0188] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0189] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. A dynamic scheduling method based on a gas business middle station is characterized in that: The push method comprises the following steps: S1. Collect and pre-process real-time data information of gas business; S2. Build a knowledge graph based on data information to associate gas business data in real time; S3, based on deep learning and feature selection technology, extract key features and train the gas business data recommendation model; S4. Generate a gas service recommendation list using a recommendation algorithm and push strategy, and push it to the user; S5. Implement a feedback mechanism to dynamically optimize the recommendation algorithm and push strategy based on user feedback.

2. The dynamic scheduling method based on the gas business middle station according to claim 1 is characterized in that: The step S1 specifically includes the following steps: S11. Gas business collection includes obtaining detailed information on gas business usage plans, usage targets, and usage arrangements, which are collected through gas business systems or online platforms; S12. Use the Pandas library in Python to clean the data, including removing duplicates, processing missing data, and unifying the format to ensure the integrity and availability of the data; S13. Perform word segmentation and part-of-speech tagging on the collected text data of gas business, extract keywords and topics using natural language processing technology, and identify content related to gas business needs; S14, perform data normalization processing to unify data from different sources and formats into a certain standardized format to facilitate subsequent feature extraction and model training; S15. Use data mining technology to analyze and mine historical gas business data to discover potential patterns and trends hidden in the data to support more accurate gas business recommendations; S16. Based on the collected gas usage data, use machine learning or rule engine technology to establish a personalized gas demand model; S17. Convert the gas usage demand description into a structured data representation so that the system can understand and process it; S18. Use text mining and sentiment analysis techniques to analyze the text descriptions of users, identify their focus, preferences and special needs for gas business, and provide guidance for subsequent recommendation systems; S19. Integrate gas data from different sources to ensure the integrity and consistency of the integrated data set; S110. Implement data verification and quality control, including data legitimacy verification, logical consistency check and abnormal data processing to ensure data reliability and validity; S111. Use visualization tools to conduct exploratory analysis on data to discover the distribution and potential associations of gas business data, providing a basis for subsequent knowledge graph construction and model training; S112. Establish data management strategies and security measures to ensure the privacy protection and information security of users and user data.

3. The dynamic scheduling method based on the gas business middle station according to claim 2 is characterized in that: The step S2 specifically includes the following steps: S21. Use natural language processing technology to semantically annotate and identify entities in gas business data, identify key concepts, topics and entities, and convert text information into structured knowledge graph data based on part-of-speech tagging and named entity recognition technology to facilitate subsequent relationship modeling and reasoning; S22. Establish a conceptual hierarchical structure of the knowledge graph, including basic information of users and basic concepts of gas usage, and perform hierarchical representation; S23. Use graph database to store and manage knowledge graph data, support complex relationship modeling and query operations, use reasoning algorithms to analyze implicit relationships and dependencies in knowledge graphs, and enhance the intelligence and recommendation accuracy of gas usage systems; S24. Design an update mechanism for the knowledge graph, regularly update and maintain course content and gas business information, and ensure the timeliness and integrity of the knowledge graph; S25. Introduce an incremental update strategy to update the gas usage map based on new user requirements and usage needs to reflect the latest user usage data; S26. Implement semantic reasoning function, analyze the logical relationship between the gas service needs of users and the relevant content in the knowledge graph through the reasoning engine, and provide support for the recommendation system; S27. Develop an efficient search engine that supports complex semantic-based queries and information retrieval to improve the efficiency and accuracy of searching for gas business information.

4. The dynamic scheduling method based on the gas business middle station according to claim 3 is characterized in that: The step S3 specifically comprises the following steps: S31. Based on deep learning technology, convolutional neural network and recurrent neural network models are used for feature extraction; S32. Design corresponding feature extractors according to different types of gas business data to identify and extract key features related to gas business needs; S33. Select the appropriate deep learning model architecture and algorithm, and configure the model according to the characteristics of the data and the requirements of the task; S34. Use TensorFlow or PyTorch framework to train and optimize the model, and adjust the model's hyperparameters and learning strategies to improve the model's generalization ability and recommendation accuracy; S35. Introduce cross-validation technology and ensemble learning methods to integrate the prediction results of multiple models to improve the stability and reliability of gas business recommendations; S36. Design evaluation indicators to evaluate the performance of the trained model; S37. Use validation sets and test sets to verify the generalization ability of the model, analyze the performance differences of the model in different scenarios, optimize model parameters and recommendation algorithms based on feedback data and user evaluation, and continuously improve the intelligence level of the system and user experience; S38. Implement the deployment and integration of the trained model and integrate it into the recommendation system to support real-time gas business recommendation services; S39. Configure the service ports and interfaces of the model to achieve rapid model calling and data interaction, ensuring the efficiency and stability of the system; S310. Design model update and upgrade strategies, monitor model operation status and performance indicators, and timely adjust and optimize system operation efficiency.

5. The dynamic scheduling method based on the gas business middle station according to claim 4 is characterized in that: The step S4 specifically comprises the following steps: S41. Implement TF-IDF, Word2Vec or Doc2Vec content-based recommendation algorithm to analyze the text content and semantic similarity of gas business information and recommend information that matches the needs of users; S42. Combining keyword extraction and text clustering technology, discover hidden associations and information in gas business data to improve the accuracy and diversity of recommendations; S43. Introduce user-based collaborative filtering algorithm to analyze the similarities and preferences between users and peers, and recommend information with similar gas business history; S44. Implement project-based collaborative filtering algorithms to analyze the association and cross-use of gas business data and provide cross-disciplinary and cross-field recommendation services; S45. Design a multi-objective optimization algorithm to balance the relevance and novelty of recommended information to meet the personalized gas service needs of users; S46. Combine fuzzy logic or evolutionary algorithms to solve the trade-off between diversity and accuracy in the recommendation process and improve the adaptability and intelligence of the system.

6. The dynamic scheduling method based on the gas business middle station according to claim 5 is characterized in that: The step S7 specifically includes the following push system implementation: S471. User interface design: Design an intuitive and friendly user interface that supports personalized user settings and operations, enables browsing, searching and selecting gas business information, and uses React or Vuejs front-end frameworks to develop responsive design interfaces that adapt to different terminal devices and improve user experience and usability; S472. Backend service development: Develop stable and efficient backend services to handle user requests and data interactions, support the invocation of recommendation algorithms and models, use Flask or Django backend frameworks, implement server-side API interfaces, support dynamic loading of data and real-time updates of recommendation results; S473, Data interface and external resource integration: Implement data interface with external gas resource library, obtain the latest gas business information and gas packages, expand the content coverage and information update speed of the system, design data synchronization and integration strategies, ensure the consistency and timeliness of data between the system and external resource library, and avoid information islands and outdated data problems; S474. System performance optimization and security protection: Optimize system performance, including optimizing the loading speed of the front-end interface and improving the concurrent processing capabilities of back-end services, ensure the stability and response speed of the system, introduce data encryption and user identity authentication mechanisms, protect the security of users and their personal information, and comply with relevant data protection laws and regulations and privacy policy requirements.

7. The dynamic scheduling method based on the gas service middle station according to claim 6 is characterized in that: The step S7 specifically includes the following steps: S51. Develop personalized recommendation strategies based on recommendation strategies, adjust recommendation results according to historical behaviors and real-time feedback of users, and provide dynamic adjustment and optimization of gas business information push services; S52. Implement user feedback collection and analysis functions based on the feedback mechanism, collect users’ evaluations and opinions on recommended materials, and provide real-time data support for recommendation algorithms and system optimization.

8. The dynamic scheduling system based on the gas business middle station according to claim 7 is characterized in that: The push system comprises: Data collection and preprocessing module, used to collect real-time data information of gas business, including data cleaning, normalization and feature extraction to ensure data accuracy and consistency; The knowledge graph construction module is used to build a knowledge graph using natural language processing technology to associate gas business data in real time. The knowledge graph includes course topics, knowledge points, user goals and related reference materials, and performs semantic annotation and relationship reasoning on them. The knowledge graph includes course topics, knowledge points, user goals and related reference materials, and performs semantic annotation and relationship reasoning on them. Feature selection and model training module, based on deep learning and feature selection technology, extracts key features related to gas business needs, uses convolutional neural network and recursive neural network for model training, and establishes a gas business data recommendation model; The recommendation algorithm and push strategy module implements a combination of content-based recommendation algorithm and collaborative filtering algorithm to generate a gas business recommendation list, using a multi-objective optimization algorithm to balance the relevance and novelty of the recommendation; Push system real module, design and implement gas business data push system, including user interface, background service and data interface; Feedback mechanism and system optimization module implement feedback mechanism, allowing users to evaluate and provide feedback on the pushed gas business information, dynamically adjust and optimize the recommendation algorithm and push strategy based on feedback data, improve the intelligence level of the system and user satisfaction, and regularly update and maintain the system to ensure its continued effectiveness and adapt to new user needs.

9. The dynamic scheduling system based on the gas business middle station according to claim 8 is characterized in that: The data collection and preprocessing module specifically includes: Data collection module: responsible for collecting data from various data sources, real-time data information of gas business; Data cleaning module, used to clean the collected data, remove noise, duplicate data, and handle missing values; Data normalization module, standardizing data formats to ensure data consistency and comparability; Feature extraction module, which extracts useful features from cleaned and normalized data for further analysis or model training; The data management and storage module manages and stores processed data to ensure data security, accessibility and continuous updating.

10. The dynamic scheduling system based on the gas business middle station according to claim 8 is characterized in that: The knowledge graph construction module specifically includes: Entity recognition and extraction module, used to identify and extract key entities in text and data information related to gas business; The relation extraction and annotation module is used to extract the relations between entities from the text and annotate the relations; The semantic annotation and reasoning module performs deep semantic annotation on entities and relationships in the knowledge graph and uses the reasoning engine to perform logical reasoning; The graph update and maintenance module maintains the integrity and accuracy of the knowledge graph and updates and expands the graph content in a timely manner; Visualization and application module: Provides visualization display and application interface of knowledge graph to facilitate users to understand and use graph information.

11. The dynamic scheduling system based on the gas business middle station according to claim 8 is characterized in that: The feature selection and model training module specifically includes: Feature selection module, used to select the key features most relevant to model training from the extracted features, reduce redundancy and improve model performance; Model building module for building deep learning models using convolutional neural networks and recurrent neural networks for data modeling; Model training and verification module: train deep learning models and verify their performance to ensure the accuracy and generalization ability of the models; Model evaluation and tuning module: evaluates the performance of the model and optimizes the model based on the evaluation results; Model deployment and application module: deploy the trained model to the production environment and integrate it into the gas business recommendation system.

Images