User behavior data mining and prediction analysis system in intelligent power grid environment

Abstract

The invention belongs to the technical field of power grids, and discloses a user behavior data mining and prediction analysis system in a smart power grid environment, which comprises a data acquisition module, a multi-dimensional feature extraction module, a behavior pattern recognition module, a demand prediction and trend analysis module, a personalized service recommendation module and a data storage and security module. According to the method, the accuracy of user behavior pattern recognition is improved, complex features can be extracted from the power consumption behaviors of the users through combination of multi-dimensional feature extraction and a deep learning algorithm, and therefore the behavior patterns of the users are recognized more accurately. Compared with simple clustering of traditional single-dimensional features, the system can recognize multi-level behavior features such as power utilization peak and valley periods and seasonal preference, and the comprehensiveness of user behavior portraits is improved. Through a demand prediction technology combining time sequence analysis and an integrated learning model, accurate short-term and long-term power consumption prediction can be carried out, and a demand change trend is judged.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grids, and in particular to a user behavior data mining and prediction analysis system in a smart grid environment. Background Art

[0002] In the smart grid environment, existing technologies widely use data mining technology to analyze users' electricity consumption behavior data, mainly in the analysis of user electricity load characteristics, power anomaly detection and simple demand forecasting. This type of system collects user power data through smart meters and uses traditional statistical models or time series predictions to perform certain feature recognition and predictions on user electricity consumption behavior. However, existing systems usually have difficulty in accurately identifying the diverse characteristics of users' electricity consumption behavior, and are insufficient in processing the complex correlations between multi-dimensional user data, which affects the accuracy of personalized services.

[0003] Through the above analysis, the problems and defects of the prior art are as follows:

[0004] (1) Insufficient data dimension: The system’s understanding of users’ electricity consumption behavior is limited to a single dimension, such as daily electricity load, which makes it difficult to fully describe the users’ complex electricity consumption characteristics.

[0005] (2) Insufficient real-time and adaptability: Existing technologies are difficult to dynamically respond to changes in user behavior and lack the ability to accurately predict and adapt to electricity consumption trends in real time.

[0006] (3) Insufficient personalized service support: Failure to provide differentiated and customized electricity service support for different users makes it difficult to achieve intelligent and customized electricity optimization distribution. Summary of the invention

[0007] In view of the problems existing in the prior art, the present invention provides a user behavior data mining and prediction analysis system in a smart grid environment.

[0008] The present invention is implemented as follows: a user behavior data mining and prediction analysis system in a smart grid environment includes:

[0009] Data collection module, multi-dimensional feature extraction module, behavior pattern recognition module, demand forecasting and trend analysis module, personalized service recommendation module, data storage and security module;

[0010] The data collection module is connected with the multi-dimensional feature extraction module and the personalized service recommendation module, and is used to collect user electricity consumption data in real time through smart meters and sensor networks, including electricity consumption, power load, electricity price fluctuations, climate data, and user behavior preferences;

[0011] The multi-dimensional feature extraction module is connected with the data acquisition module, the behavior pattern recognition module, and the data storage and security module, and is used to extract feature data that can characterize the user's electricity consumption behavior from different dimensions, covering time, space, behavior preferences, and environmental factors, by using data cleaning, preprocessing, and normalization methods;

[0012] The behavior pattern recognition module is connected with the multi-dimensional feature extraction module and the demand forecasting and trend analysis module to use deep learning and clustering algorithms to identify the patterns and habits of users' electricity consumption behaviors and form user behavior portraits;

[0013] The demand forecasting and trend analysis module is connected with the behavior pattern recognition module, the personalized service recommendation module, and the data storage and security module to forecast electricity demand through the time series analysis model and the prediction algorithm based on ensemble learning, and output personalized load forecasting results;

[0014] The personalized service recommendation module is connected to the demand forecasting and trend analysis module to generate personalized electricity usage suggestions for different users based on user behavior profiles and demand forecasts. Power supply companies can provide adaptive electricity price packages and energy-saving measures accordingly.

[0015] The data storage and security module is connected with the data acquisition module, the multi-dimensional feature extraction module, and the demand forecasting and trend analysis module, and is used to use distributed database and encryption storage technology to ensure efficient storage, query and security of data.

[0016] Further, the data acquisition module implementation method:

[0017] Equipment installation: Install smart meters and various sensors at the user end, including current and voltage sensors, temperature and humidity sensors;

[0018] Real-time data collection: Smart meters are used to record users’ electricity consumption, voltage, and power load data in real time, and sensors are used to collect environmental and climate data. IoT technology is used to send the collected data to a centralized data collection platform through wireless transmission protocols including ZigBee and Wi-Fi.

[0019] Multi-source data integration: In the integration platform, the received data is merged and timestamped to generate a data set that can be analyzed.

[0020] Further, the multi-dimensional feature extraction module implementation method:

[0021] Data cleaning and denoising: outliers in the data are processed through outlier processing and interpolation methods to remove invalid or distorted data;

[0022] Feature normalization: Normalize data from different sources, normalize power data and environmental data to the same numerical range, so that they are consistent in subsequent analysis;

[0023] Feature engineering: Combine domain knowledge and statistical analysis to generate feature vectors, including peak and valley electricity consumption characteristics, and the relationship between temperature, humidity and electricity consumption; use PCA (principal component analysis) or t-SNE (t-distributed stochastic neighbor embedding) dimensionality reduction methods to further simplify the data, reducing the data dimension while retaining key features.

[0024] Further, the behavior pattern recognition module implementation method:

[0025] Data input: The feature data generated by the multi-dimensional feature extraction module is used as input;

[0026] Cluster analysis: Use K-means or DBSCAN clustering algorithms to group user data and identify user groups with similar electricity usage patterns, such as those with high peak electricity usage and large seasonal electricity usage variations.

[0027] Deep learning model: Use LSTM long short-term memory network or convolutional neural network CNN to perform time series analysis on the user's historical electricity consumption data, thereby capturing the user's long-term and short-term electricity consumption behavior;

[0028] Behavioral profile generation: The model outputs each user's behavioral profile, including electricity usage patterns, electricity usage preferences, and time period preferences, for subsequent demand forecasting and personalized service use.

[0029] Further, the demand forecasting and trend analysis module is implemented as follows:

[0030] Time Series Forecasting Models:

[0031] A time series data set is constructed based on historical electricity consumption data, and the ARIMA model is used to predict the short-term electricity demand of a single user or a single cell;

[0032] For long-term data, the LSTM model is used to forecast user electricity demand on a daily, weekly or monthly basis to take into account the changing trends in the long-term time series;

[0033] Ensemble learning algorithm:

[0034] Combining random forest and XGBoost algorithms, the prediction results of multiple models are integrated and optimized to ensure the stability and prediction accuracy of the model in complex scenarios;

[0035] The integrated model can accurately predict future electricity demand under various characteristic conditions, enabling the system to respond in real time under different weather and user behavior changes;

[0036] Trend analysis: Conduct data trend analysis on the forecast results to identify changes in users' electricity demand, including seasonal fluctuations and temporary peak electricity demand.

[0037] Further, the personalized service recommendation module implementation method:

[0038] Recommendation algorithm: Use collaborative filtering recommendation algorithm, combined with user portrait data and demand prediction results, to push personalized recommendations to different users;

[0039] Content-based recommendations: Analyze users’ electricity usage habits and behavior profiles, and recommend the best electricity packages and time strategies;

[0040] Recommendation based on collaborative filtering: Provide group recommendations for users with similar electricity usage patterns;

[0041] Personalized suggestion generation: By analyzing each user's behavior pattern and load forecast results, different energy-saving plans are pushed to users, including electricity usage suggestions for specific time periods and usage suggestions during peak electricity price periods;

[0042] User feedback collection: The system supports collecting user feedback and behavior change data, continuously optimizing recommendation algorithms, and improving user experience;

[0043] Data storage and security module implementation method:

[0044] Distributed database: Use distributed database to store massive user data to ensure efficient reading, writing and querying of data; distributed storage can process large amounts of real-time data and support horizontal expansion;

[0045] Encrypted storage: During data storage and transmission, AES and SSL / TLS protocols are used to encrypt user data to ensure user privacy and security;

[0046] Access control: Through the role-based access control mechanism, ensure that users with different permissions can only access data with corresponding permissions, thereby avoiding the leakage of user data.

[0047] Another object of the present invention is to provide a user behavior data mining and prediction analysis method in a smart grid environment, comprising:

[0048] Step 1: Use smart meters and sensor networks through the data collection module to collect user electricity consumption data in real time, including electricity consumption, power load, electricity price fluctuations, climate data, and user behavior preferences;

[0049] Step 2: Using data cleaning, preprocessing and normalization methods through the multidimensional feature extraction module, feature data that can characterize the user's electricity consumption behavior is extracted from different dimensions, covering time, space, behavior preferences and environmental factors;

[0050] Step 3: Use deep learning and clustering algorithms through the behavior pattern recognition module to identify the patterns and habits of the user's electricity usage behavior and form a user behavior profile;

[0051] Step 4: Use the demand forecasting and trend analysis module to forecast electricity demand using the time series analysis model and the prediction algorithm based on ensemble learning, and output personalized load forecast results;

[0052] Step 5: Generate personalized electricity usage suggestions for different users based on user behavior profiles and demand forecasts through the personalized service recommendation module, and power supply companies can provide adaptive electricity price packages and energy-saving measures accordingly;

[0053] Step 6: Use distributed database and encrypted storage technology through the data storage and security module to ensure efficient storage, query and security of data.

[0054] Another object of the present invention is to provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the user behavior data mining and predictive analysis method in the smart grid environment.

[0055] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to perform the steps of the user behavior data mining and predictive analysis method in the smart grid environment.

[0056] Another object of the present invention is to provide an information data processing terminal, which is used to implement the user behavior data mining and predictive analysis system in the smart grid environment.

[0057] In combination with the above technical solutions and the technical problems solved, please analyze the advantages and positive effects of the technical solutions to be protected by the present invention from the following aspects:

[0058] First, in view of the technical problems existing in the above-mentioned prior art and the difficulty of solving the problems, the technical solutions to be protected by the present invention and the results and data during the research and development process are closely combined to analyze in detail and deeply how the technical solutions of the present invention solve the technical problems, and some creative technical effects brought about after solving the problems. The specific description is as follows:

[0059] 1. Improved the accuracy of user behavior pattern recognition:

[0060] The combination of multi-dimensional feature extraction and deep learning algorithms enables the system to extract complex features from users' electricity consumption behaviors, thereby more accurately identifying users' behavior patterns. Compared with the simple clustering of traditional single-dimensional features, the system can identify multi-level behavioral features such as peak and valley periods of electricity consumption and seasonal preferences, significantly improving the comprehensiveness of user behavior portraits.

[0061] 2. Achieve high-precision electricity demand forecasting:

[0062] Through demand forecasting technology that combines time series analysis with integrated learning models, the system can make accurate short-term and long-term electricity forecasts and judge demand trends. Improved forecast accuracy helps power suppliers optimize resource allocation, thereby avoiding power waste or insufficient supply.

[0063] 3. Meet personalized service needs:

[0064] The system generates personalized electricity usage recommendations based on user profiles, enabling power suppliers to provide differentiated services to meet the electricity usage preferences of different users. Personalized services not only improve user experience, but also encourage users to actively adjust their electricity usage behaviors to achieve the goals of energy saving and smart electricity usage.

[0065] 4. Improved the real-time adaptability of the system:

[0066] The real-time monitoring and feedback module realizes the dynamic adaptability of the system. The system can respond to changes in user behavior in real time and adjust the prediction model through the model's feedback mechanism, thereby continuously optimizing the prediction effect and the effectiveness of personalized services.

[0067] 5. Enhanced data security and privacy protection:

[0068] The system's distributed storage and encryption technology provides higher security for user data and reduces the risk of data leakage. Data security helps to enhance user trust and make users more willing to share behavioral data, thus forming a data closed loop and further optimizing system performance.

[0069] Through the integrated application of multi-dimensional feature extraction, behavior pattern recognition, demand forecasting and personalized service recommendation, this system not only provides advanced solutions for user electricity behavior data mining and predictive analysis, but also significantly improves the service capabilities of power suppliers and user experience. The system specifically solves the problems of traditional systems' difficulty in identifying user diversified behaviors, low prediction accuracy and insufficient personalized services in a real-time adaptive manner, and promotes data-driven innovation and development in the field of smart grids.

[0070] Second, in the smart grid environment, the user behavior data mining and prediction analysis system achieves a series of specific technical effects through multi-module collaborative processing, which are mainly manifested in the following aspects:

[0071] 1. Real-time data collection and multi-dimensional feature extraction: Through smart meters and sensor networks, the system can collect multi-dimensional data such as users' electricity consumption, power load, electricity price fluctuations, climate data, and behavioral preferences in real time. For example, in the daily electricity consumption data collection of a family, the system records the hourly electricity consumption fluctuations in real time, and combines the temperature and humidity data fed back by the climate sensor to make the multi-dimensional features of the user behavior data more complete. In this way, the multi-dimensional feature extraction module can accurately clean and standardize the data, extract features such as the user's daily peak electricity consumption, weekend and weekday electricity consumption patterns, and lay the foundation for the next step of identifying electricity consumption behavior.

[0072] 2. High-precision behavioral pattern recognition: Through the combination of deep learning models and clustering algorithms, the system identifies the user's electricity usage habits and forms a behavioral portrait of the user. For example, through the clustering algorithm, the system found that a certain user's electricity consumption increased significantly from 8 to 10 o'clock every night, and the proportion of air-conditioning load in summer increased to more than 40% of the daily average total load. Based on these characteristics, the behavioral pattern recognition module generates a characteristic portrait of the user, including labels such as "peak electricity consumption at night" and "high air-conditioning load in summer", providing a basis for subsequent personalized recommendations. This deep recognition based on user habits improves the accuracy of user portraits and enables power supply to more effectively adapt to user needs.

[0073] 3. Accurate demand forecasting and trend analysis: When predicting electricity demand, the system uses time series analysis models and integrated learning algorithms to predict future electricity demand with high accuracy. For example, by analyzing the electricity consumption data of users in a certain community during the summer peak period in the past five years, the system predicts that the load growth rate during the future summer peak period will reach 5-8%. This load forecasting result not only improves the dispatching efficiency of the power grid, but also helps power supply companies predict the peak load period in advance, thereby optimizing the allocation of power resources.

[0074] 4. Accuracy of personalized service recommendations: Based on user behavior portraits and demand forecasts, the system can generate personalized electricity usage recommendations for different users, significantly improving the user's electricity efficiency. For example, the system identifies that a user's electricity consumption is concentrated in low-peak hours, and recommends the user to choose the "night low-price package", which is expected to reduce its monthly electricity bill by 15%. For users whose electricity consumption fluctuates frequently, the system recommends intelligent electricity management suggestions to help users reduce peak loads and fill valley loads, reducing energy waste caused by load fluctuations. These personalized suggestions effectively help users optimize their electricity usage structure while saving electricity bills.

[0075] 5. Efficient data storage and security assurance: Throughout the entire data processing process, the system uses distributed database and encrypted storage technology to ensure that data is efficient and secure during transmission and storage. For example, all user behavior data and prediction data are stored in an encrypted database. Distributed storage speeds up data query and provides permission management when accessing data. This not only protects data privacy, but also provides reliable data support for power supply companies, making the data management of the entire system more secure and scalable.

[0076] Through the coordinated application of the above modules and data processing methods, the system achieves multiple technical effects such as real-time and accurate monitoring, behavior recognition, demand forecasting, personalized recommendations and data security, which improves the grid's ability to respond to users' electricity demand and provides users with accurate and efficient smart electricity support. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 It is a structural block diagram of a user behavior data mining and predictive analysis system in a smart grid environment provided by an embodiment of the present invention.

[0078] Figure 2 It is a flow chart of the data acquisition module implementation method provided by an embodiment of the present invention.

[0079] Figure 3 It is a flow chart of a method for user behavior data mining and predictive analysis in a smart grid environment provided by an embodiment of the present invention.

[0080] Figure 1 In: 1. Data acquisition module; 2. Multi-dimensional feature extraction module; 3. Behavior pattern recognition module; 4. Demand forecasting and trend analysis module; 5. Personalized service recommendation module; 6. Data storage and security module. DETAILED DESCRIPTION

[0081] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0082] In the smart grid environment, the user behavior data mining and prediction analysis system first collects various electricity-related data of users in real time through the data acquisition module, relying on smart meters and sensor networks. The collected data includes users' real-time electricity consumption, power load changes, electricity price fluctuations, climate conditions, and user preference behavior data. The acquisition module converts the original data stream into the standardized data input required by the system for subsequent processing and analysis.

[0083] Next, the data enters the multidimensional feature extraction module for deep preprocessing. This module will first perform data cleaning to remove noise and outliers to ensure data quality. Then, normalization and standardization techniques are used to convert the data to a unified dimension range to eliminate the dimensional differences between different data sources. The extracted feature data constructs a multidimensional feature matrix that can accurately characterize the user's electricity consumption behavior from the dimensions of time, space, behavioral preferences, and environmental factors, providing a data basis for the identification of user behavior patterns.

[0084] In the behavior pattern recognition module, the system uses deep learning and clustering algorithms to mine user feature data. Based on deep neural networks, the module can automatically extract potential behavior patterns, group user behaviors through clustering analysis, and identify individual users' electricity usage habits and patterns. For example, the system can identify users' daily peak electricity consumption time, seasonal load change characteristics, and changes in electricity consumption behavior under specific weather conditions to form a behavioral portrait of each user.

[0085] The behavior profile and user power consumption pattern information will be further transmitted to the demand forecasting and trend analysis module, which will use the time series model and integrated learning algorithm to predict the user's future power demand. This module will use historical power consumption data for trend analysis, predict the user's power load fluctuations in different periods, and generate load demand forecast results. The forecast results are further processed into personalized load data to provide support for users' future power consumption.

[0086] Finally, the personalized service recommendation module generates user-specific electricity usage recommendations based on the predicted electricity demand and user behavior profile. For example, the system can recommend users to choose electricity price packages that are more in line with their electricity usage patterns, or provide specific energy-saving recommendations to help users optimize their electricity usage structure. The data storage and security module runs through the entire process, using an encrypted distributed database for storage and query management of all data to ensure efficient data storage and privacy security.

[0087] like Figure 1 As shown, a user behavior data mining and prediction analysis system in a smart grid environment provided by an embodiment of the present invention includes:

[0088] Data collection module 1, multi-dimensional feature extraction module 2, behavior pattern recognition module 3, demand forecasting and trend analysis module 4, personalized service recommendation module 5, data storage and security module 6.

[0089] The data collection module 1 is connected to the multi-dimensional feature extraction module 2 and the personalized service recommendation module 5, and is used to collect user electricity consumption data in real time through smart meters and sensor networks, including electricity consumption, power load, electricity price fluctuations, climate data, and user behavior preferences;

[0090] The multi-dimensional feature extraction module 2 is connected to the data acquisition module 1, the behavior pattern recognition module 3, and the data storage and security module 6, and is used to extract feature data that can characterize the user's electricity consumption behavior from different dimensions, covering time, space, behavior preference and environmental factors, by using data cleaning, preprocessing and normalization methods;

[0091] The behavior pattern recognition module 3 is connected with the multi-dimensional feature extraction module 2 and the demand prediction and trend analysis module 4, and is used to use deep learning and clustering algorithms to identify the patterns and habits of the user's electricity consumption behavior and form a user behavior portrait;

[0092] The demand forecasting and trend analysis module 4 is connected with the behavior pattern recognition module 3, the personalized service recommendation module 5, and the data storage and security module 6, and is used to forecast the power demand through the time series analysis model and the prediction algorithm based on ensemble learning, and output the personalized load forecasting result;

[0093] The personalized service recommendation module 5 is connected to the demand forecasting and trend analysis module 4 and is used to generate personalized electricity usage suggestions for different users based on user behavior profiles and demand forecasts, so that power supply companies can provide adaptive electricity price packages and energy-saving measures suggestions accordingly;

[0094] The data storage and security module 6 is connected to the data acquisition module 1, the multi-dimensional feature extraction module 2, and the demand forecasting and trend analysis module 4, and is used to use distributed database and encryption storage technology to ensure efficient storage, query and security of data.

[0095] like Figure 2 As shown, the data acquisition module implementation method provided by the embodiment of the present invention is:

[0096] S101, equipment installation: install smart meters and various sensors at the user end, including current and voltage sensors, temperature and humidity sensors;

[0097] S102, real-time data collection:

[0098] Smart meters are used to record users' electricity consumption, voltage, and power load data in real time, and sensors are used to collect environmental and climate data. IoT technology is used to send the collected data to a centralized data collection platform through wireless transmission protocols including ZigBee and Wi-Fi.

[0099] S103, Multi-source data integration:

[0100] In the integration platform, the received data is merged and labeled with time stamps to generate a data set that can be analyzed.

[0101] The multi-dimensional feature extraction module implementation method provided by the embodiment of the present invention is as follows:

[0102] Data cleaning and denoising: outliers in the data are processed through outlier processing and interpolation methods to remove invalid or distorted data;

[0103] Feature normalization: Normalize data from different sources, normalize power data and environmental data to the same numerical range, so that they are consistent in subsequent analysis;

[0104] Feature engineering: Combine domain knowledge and statistical analysis to generate feature vectors, including peak and valley electricity consumption characteristics, and the relationship between temperature, humidity and electricity consumption; use PCA (principal component analysis) or t-SNE (t-distributed stochastic neighbor embedding) dimensionality reduction methods to further simplify the data, reducing the data dimension while retaining key features.

[0105] The behavior pattern recognition module implementation method provided by the embodiment of the present invention is as follows:

[0106] Data input: The feature data generated by the multi-dimensional feature extraction module is used as input;

[0107] Cluster analysis: Use K-means or DBSCAN clustering algorithms to group user data and identify user groups with similar electricity usage patterns, such as those with high peak electricity usage and large seasonal electricity usage variations.

[0108] Deep learning model: Use LSTM long short-term memory network or convolutional neural network CNN to perform time series analysis on the user's historical electricity consumption data, thereby capturing the user's long-term and short-term electricity consumption behavior;

[0109] Behavioral profile generation: The model outputs each user's behavioral profile, including electricity usage patterns, electricity usage preferences, and time period preferences, for subsequent demand forecasting and personalized service use.

[0110] The implementation method of the demand forecasting and trend analysis module provided by the embodiment of the present invention is as follows:

[0111] Time Series Forecasting Models:

[0112] A time series data set is constructed based on historical electricity consumption data, and the ARIMA model is used to predict the short-term electricity demand of a single user or a single cell;

[0113] For long-term data, the LSTM model is used to forecast user electricity demand on a daily, weekly or monthly basis to take into account the changing trends in the long-term time series;

[0114] Ensemble learning algorithm:

[0115] Combining random forest and XGBoost algorithms, the prediction results of multiple models are integrated and optimized to ensure the stability and prediction accuracy of the model in complex scenarios;

[0116] The integrated model can accurately predict future electricity demand under various characteristic conditions, enabling the system to respond in real time under different weather and user behavior changes;

[0117] Trend analysis: Conduct data trend analysis on the forecast results to identify changes in users' electricity demand, including seasonal fluctuations and temporary peak electricity demand.

[0118] The implementation method of the personalized service recommendation module provided by the embodiment of the present invention is as follows:

[0119] Recommendation algorithm: Use collaborative filtering recommendation algorithm, combined with user portrait data and demand prediction results, to push personalized recommendations to different users;

[0120] Content-based recommendations: Analyze users’ electricity usage habits and behavior profiles, and recommend the best electricity packages and time strategies;

[0121] Recommendation based on collaborative filtering: Provide group recommendations for users with similar electricity usage patterns;

[0122] Personalized suggestion generation: By analyzing each user's behavior pattern and load forecast results, different energy-saving plans are pushed to users, including electricity usage suggestions for specific time periods and usage suggestions during peak electricity price periods;

[0123] User feedback collection: The system supports collecting user feedback and behavior change data, continuously optimizing recommendation algorithms, and improving user experience;

[0124] Data storage and security module implementation method:

[0125] Distributed database: Use distributed database to store massive user data to ensure efficient reading, writing and querying of data; distributed storage can process large amounts of real-time data and support horizontal expansion;

[0126] Encrypted storage: During data storage and transmission, AES and SSL / TLS protocols are used to encrypt user data to ensure user privacy and security;

[0127] Access control: Through the role-based access control mechanism, ensure that users with different permissions can only access data with corresponding permissions, thereby avoiding the leakage of user data.

[0128] like Figure 3 As shown, a user behavior data mining and prediction analysis method in a smart grid environment provided by an embodiment of the present invention includes:

[0129] S201, using a data collection module to use smart meters and sensor networks to collect user electricity consumption data in real time, including electricity consumption, power load, electricity price fluctuations, climate data, and user behavior preferences;

[0130] S202, extracting feature data capable of characterizing the user's electricity consumption behavior from different dimensions, including time, space, behavior preference, and environmental factors, by using a multi-dimensional feature extraction module using data cleaning, preprocessing, and normalization methods;

[0131] S203, using a behavior pattern recognition module to use deep learning and clustering algorithms to identify patterns and habits of the user's electricity usage behavior and form a user behavior profile;

[0132] S204, using the demand forecasting and trend analysis module to forecast electricity demand using a time series analysis model and a forecasting algorithm based on ensemble learning, and outputting a personalized load forecasting result;

[0133] S205, generating personalized electricity usage suggestions for different users based on user behavior profiles and demand predictions through a personalized service recommendation module, and power supply companies can provide adaptive electricity price packages and energy-saving measures suggestions accordingly;

[0134] S206, using distributed database and encryption storage technology through the data storage and security module to ensure efficient storage, query and security of data.

[0135] Another object of the present invention is to provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the user behavior data mining and predictive analysis method in the smart grid environment.

[0136] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to perform the steps of the user behavior data mining and predictive analysis method in the smart grid environment.

[0137] Another object of the present invention is to provide an information data processing terminal, which is used to implement the user behavior data mining and predictive analysis system in the smart grid environment.

[0138] The present invention is specifically implemented:

[0139] This system integrates a variety of data processing and analysis technologies for user behavior data of smart grids. The following are the specific implementation methods of each module:

[0140] 1. Data acquisition module

[0141] Implementation method:

[0142] Equipment installation: Install smart meters and various sensors at the user end, such as current and voltage sensors, temperature and humidity sensors.

[0143] Real-time data collection: Smart meters record users’ electricity consumption, voltage, power load and other data in real time, and sensors collect environmental and climate data. Using the Internet of Things (IoT) technology, the collected data is sent to a centralized data collection platform via wireless transmission protocols (such as ZigBee and Wi-Fi).

[0144] Multi-source data integration: In the integration platform, the received data is merged and labeled with timestamps to generate a data set that can be analyzed.

[0145] Data Type:

[0146] 1. Electricity consumption data: including daily electricity consumption, voltage, and current change curves.

[0147] 2. Meteorological data: including environmental parameters such as temperature, humidity, and rainfall, which are used to identify seasonal changes.

[0148] 3. User behavior data: collects information such as appliance usage time and load changes to reflect the user's specific electricity usage behavior patterns.

[0149] 2. Multidimensional feature extraction module

[0150] Implementation method:

[0151] Data cleaning and denoising: outliers in the data are processed through outlier processing, interpolation methods, etc. to remove invalid or distorted data.

[0152] Feature normalization: Normalize data from different sources, normalize power data, environmental data, etc. to the same numerical range to ensure consistency in subsequent analysis.

[0153] Feature engineering: Combine domain knowledge and statistical analysis to generate feature vectors, such as peak and valley electricity consumption characteristics, and the relationship between temperature, humidity and electricity consumption. Use dimensionality reduction methods such as PCA (principal component analysis) or t-SNE (t-distributed stochastic neighbor embedding) to further simplify the data, reducing the data dimension while retaining key features.

[0154] Feature Type:

[0155] 1. Time characteristics: daily / weekly / seasonal electricity consumption characteristics.

[0156] 2. Spatial characteristics: Differences in electricity consumption patterns in different geographical areas.

[0157] 3. Behavioral characteristics: frequency of use and time preference of specific electrical appliances.

[0158] 4. Environmental characteristics: the impact of changes in ambient temperature and humidity on electricity consumption.

[0159] 3. Behavior pattern recognition module

[0160] Implementation method:

[0161] Data input: The feature data generated by the multidimensional feature extraction module is used as input.

[0162] Cluster analysis: Use K-means or DBSCAN clustering algorithms to group user data and identify user groups with similar electricity usage patterns, such as those with high peak electricity usage and large seasonal electricity usage variations.

[0163] Deep learning model: Use LSTM (Long Short-Term Memory Network) or Convolutional Neural Network (CNN) to perform time series analysis on the user's historical electricity consumption data to capture the user's long-term and short-term electricity consumption behavior.

[0164] Behavioral profile generation: The model outputs each user's behavioral profile, including electricity usage patterns, electricity usage preferences (such as time period preferences), etc., for subsequent demand forecasting and personalized service use.

[0165] 4. Demand forecasting and trend analysis module

[0166] Implementation method:

[0167] Time Series Forecasting Models:

[0168] A time series dataset is constructed based on historical electricity consumption data, and the ARIMA model is used to predict the short-term electricity demand of a single user or a single cell.

[0169] For long-period data, the LSTM model is used to forecast user electricity demand on a daily, weekly or monthly basis to take into account the changing trends in long-term time series.

[0170] Ensemble learning algorithm:

[0171] Combining the random forest and XGBoost algorithms, the prediction results of multiple models are integrated and optimized to ensure the stability and prediction accuracy of the model in complex scenarios.

[0172] The integrated model can accurately predict future electricity demand under a variety of characteristic conditions, enabling the system to respond in real time under different weather and user behavior changes.

[0173] Trend analysis: Conduct data trend analysis on the forecast results to identify changes in users' electricity demand, such as seasonal fluctuations and temporary peak electricity demand.

[0174] 5. Personalized service recommendation module

[0175] Implementation method:

[0176] Recommendation algorithm: Use collaborative filtering recommendation algorithm, combined with user portrait data and demand prediction results, to push personalized recommendations to different users.

[0177] Content-based recommendations: Analyze users’ electricity usage habits and behavior profiles, and push the best electricity packages and time strategies.

[0178] Recommendation based on collaborative filtering: Provide group recommendations for users with similar electricity usage patterns.

[0179] Personalized suggestion generation: By analyzing each user's behavior pattern and load forecast results, different energy-saving plans are pushed to users, such as electricity usage suggestions for specific time periods and usage suggestions during peak electricity price periods.

[0180] User feedback collection: The system supports collecting user feedback and behavior change data, continuously optimizing recommendation algorithms, and improving user experience.

[0181] Sample recommendations:

[0182] 1. Optimize electricity usage time: provide reminders for low electricity price periods.

[0183] 2. Customized electricity price packages: Recommend the best electricity price plan based on the user's electricity consumption and preferences.

[0184] 3. Energy-saving suggestions: remind users to reduce unnecessary electricity consumption during peak hours to help users save electricity costs.

[0185] 6.Data storage and security module

[0186] Implementation method:

[0187] Distributed database: Use distributed databases (such as Hadoop and HBase) to store massive amounts of user data to ensure efficient reading, writing and querying of data. Distributed storage can process large amounts of real-time data and supports horizontal expansion.

[0188] Encrypted storage: During data storage and transmission, AES (Advanced Encryption Standard) and SSL / TLS protocols are used to encrypt user data to ensure user privacy and security.

[0189] Access control: Through the role-based access control (RBAC) mechanism, ensure that users with different permissions can only access data with corresponding permissions, thereby preventing the leakage of user data.

[0190] Data security measures:

[0191] 1. Encrypted transmission: Encryption protocol is used during data transmission.

[0192] 2. Multi-factor authentication: Multi-factor authentication mechanism is used when system administrators and users log in.

[0193] 3. Regular backup: Back up data regularly to prevent data loss.

[0194] Significant technological advancement brought by the system

[0195] 1. Accurate user behavior identification: Through multi-dimensional feature extraction and deep learning algorithms, the system can identify user electricity usage patterns with high accuracy, providing a reliable basis for power suppliers to optimize power distribution.

[0196] 2. Efficient electricity demand forecasting: The use of time series forecasting and integrated learning technology effectively improves the accuracy of electricity demand forecasting, helping the power grid to accurately respond to electricity load during peak periods.

[0197] 3. Personalized services and enhanced user participation: The system pushes personalized energy-saving suggestions and supports user feedback collection, so that users can not only obtain power-saving support but also actively participate in power optimization.

[0198] 4. Real-time data processing and security assurance: With the help of distributed storage and multi-layer encryption technology, the system can process large amounts of real-time data while ensuring the integrity and security of the data and enhancing user trust.

[0199] The implementation method of the system is as follows:

[0200] 1. Data Collection

[0201] Specific implementation:

[0202] Smart meters and sensor networks: Smart meters and environmental monitoring sensors are installed at each user end. Smart meters are responsible for collecting real-time power data such as voltage, current, and power consumption, while environmental sensors are responsible for recording environmental information such as temperature and humidity. In addition, smart meters can also indirectly reflect the user's power equipment usage status through power characteristics, such as power load characteristics, equipment startup and shutdown times, etc.

[0203] Data transmission protocol: Through the IoT protocol (such as ZigBee, LoRa or Wi-Fi), sensors and meters transmit the collected data to the data receiving module of the system.

[0204] Data storage: After receiving the data, it is stored in a distributed database (such as Hadoop or MongoDB) to provide efficient data access and fast query capabilities, providing the original data foundation for subsequent mining and analysis.

[0205] 2. Multidimensional feature extraction

[0206] Specific implementation:

[0207] Data cleaning: Use data cleaning algorithms (such as outlier detection and missing value filling) to remove outliers and invalid data in power data to improve data accuracy and consistency.

[0208] Data denoising: Use denoising methods such as low-pass filtering to remove unnecessary fluctuations in the collected data, improve data stability, and ensure the effectiveness of subsequent feature analysis.

[0209] Feature normalization: Normalize data of different dimensions to bring the values ​​to the same level for subsequent calculations. Use the normalization formula:

[0210] Ensure that the data is on the same scale.

[0211] Feature selection and classification: Based on the analysis requirements, the data is divided into time features (such as daily, weekly, and monthly electricity consumption data), spatial features (electricity consumption habits in different locations), behavioral features (frequency of equipment use, time period preferences), etc. Dimensionality reduction: Use algorithms such as principal component analysis (PCA) or t-distributed neighbor embedding (-SNE) to reduce the dimensionality of high-dimensional data and generate simplified feature vectors. This step can effectively reduce the computational burden while retaining the main features and improving analysis efficiency.

[0212] Feature selection and classification: Based on analysis requirements, data is divided into time features (such as daily, weekly, and monthly electricity consumption data), spatial features (electricity consumption habits in different locations), behavioral features (equipment usage frequency, time period preference), etc.

[0213] Dimensionality reduction: Use algorithms such as principal component analysis (PCA) or t-distributed neighbor embedding (t-SNE) to reduce the dimensionality of high-dimensional data and generate simplified feature vectors. This step can effectively reduce the computational burden while retaining the main features and improving analysis efficiency.

[0214] 3. Behavioral pattern recognition

[0215] Specific implementation:

[0216] Cluster analysis: Use K-means or DBSCAN clustering algorithms to perform cluster analysis on users’ electricity consumption data. Each cluster represents a typical electricity consumption behavior pattern, such as peak consumption, seasonal fluctuations, etc.

[0217] Deep learning model: Introducing the LSTM (Long Short-Term Memory) model to capture the dynamic patterns of users' electricity usage behavior over time. The LSTM model is good at processing time series data and can effectively identify long-term electricity usage habits. For example, the daily electricity usage curve of each user is analyzed to find out the seasonal and holiday electricity usage patterns.

[0218] Behavior profile generation: Based on clustering and deep learning analysis results, a behavior profile is generated for each user, recording their peak power consumption hours, device preferences, and power consumption patterns. These profiles will be used for predictive analysis and personalized recommendations.

[0219] 4. Demand forecasting and trend analysis

[0220] Specific implementation:

[0221] Time series forecasting model: The ARIMA model is used to forecast short-term electricity consumption data, which is suitable for detecting short-term demand changes on a daily or weekly basis. For long-term trends, the LSTM model is used to forecast load demand based on the user's historical electricity consumption behavior and predict electricity consumption trends for a period of time in the future.

[0222] Trend analysis: Through correlation analysis, the predicted user electricity demand is linked to their behavior profile. For example, if a certain type of user has a significant increase in electricity consumption during the peak period of summer, this trend can be identified in advance.

[0223] Ensemble learning algorithm: Use ensemble learning methods such as random forest or XGBoost to weight the results of multiple prediction models and optimize the accuracy of the final prediction. Ensemble algorithms can reduce the bias and error of a single model, thereby improving the overall prediction effect.

[0224] 5. Personalized service recommendations

[0225] Specific implementation:

[0226] Recommendation algorithm selection: Use collaborative filtering or content-based recommendation methods to provide personalized services to users.

[0227] Collaborative filtering: Analyze the electricity price packages and energy-saving habits of users with similar electricity usage behaviors, and recommend similar electricity usage plans to other users.

[0228] Content-based recommendation: By analyzing the user's behavior profile, the most suitable electricity usage strategy is recommended.

[0229] Personalized recommendation generation: The system generates recommendations for users based on user profiles and electricity consumption forecasts. For example, it recommends low-price packages for users who use a lot of electricity at night, or reminds users to reduce air conditioning use when the temperature is high.

[0230] Real-time push and feedback mechanism: Recommendations are pushed to users through user terminal applications (such as APP), and users can provide feedback on their acceptance or actual electricity usage adjustments. The system dynamically optimizes recommended content based on user feedback.

[0231] 6. Real-time monitoring and feedback

[0232] Specific implementation:

[0233] Real-time data monitoring: Upload users’ electricity consumption data to the central server in real time through wireless communication modules (such as Wi-Fi, 4G / 5G) to realize instant data collection and processing.

[0234] System feedback mechanism: Feedback prediction results and personalized recommendations to users, and feedback overall trend data to power suppliers. For example, power companies can adjust power supply loads based on real-time power consumption, or arrange power saving plans for different users.

[0235] Dynamic model update: The system regularly obtains user feedback and retrains the prediction model, dynamically adjusting model parameters to ensure that the system is adaptive to changes in user behavior, improving the accuracy of predictions and the personalization of recommendations.

[0236] In the present invention, the demand forecasting and trend analysis module uses a time series forecasting model to predict the user's future electricity demand, and establishes a personalized load forecasting model for the user through an integrated learning algorithm, wherein the forecast calculation formula for electricity demand is:

[0237] P(t)=α·f1(t)+β·f2(t)+γ·f3(t)

[0238] in:

[0239] P(t) is the predicted electricity demand at time t;

[0240] f1(t), f2(t), and f3(t) are functions of historical electricity consumption, climate factors, and electricity price fluctuations, respectively;

[0241] α, β, γ are weight coefficients obtained through training of the ensemble learning algorithm, satisfying α+β+γ=1.

[0242] 3. The user behavior data mining and prediction analysis system in the smart grid environment according to claim 1, characterized in that the data acquisition module is implemented by:

[0243] 1. Specific application fields and related products of the present invention

[0244] The user behavior data mining and prediction analysis system of the present invention is mainly used in the fields of user behavior analysis, power load management and demand prediction in the smart grid environment. Its application fields are wide, including but not limited to:

[0245] 1. Smart grid management system: Power supply companies can use this system to analyze user electricity consumption behavior and demand forecast, and optimize grid resource allocation through user behavior patterns to achieve intelligent scheduling and power load management.

[0246] 2. Smart home energy management: By integrating the system into the smart home platform, the smart home system can provide users with personalized electricity usage suggestions based on predicted electricity demand and electricity price fluctuations to improve electricity efficiency.

[0247] 3. Smart city and energy dispatch system: Through the system's demand forecasting and behavior analysis functions, the government and energy management departments can conduct city-level electricity consumption forecasting and power distribution to improve energy efficiency.

[0248] 4. Dynamic pricing in the electricity market: Power companies provide users with customized electricity price packages through predictive analysis and personalized recommendation modules, forming intelligent pricing plans to achieve user power load balance and grid stability.

[0249] 5. Energy-saving optimization service: The system can provide personalized energy-saving suggestions for industrial and commercial users, help users adjust loads, reduce peak loads and fill valleys, achieve energy conservation and emission reduction, and reduce energy costs.

[0250] 6. Distributed energy and microgrid management: For households and businesses connected to microgrids, the system can combine the power supply characteristics of distributed energy to provide personalized electricity usage guidance and achieve optimal utilization of distributed power sources.

[0251] II. Evidence of the Technical Effects Obtained by the Embodiments of the Present Invention

[0252] 1. Improve the accuracy and comprehensiveness of user electricity usage behavior data

[0253] Through the smart meter and sensor network of the data acquisition module, multi-dimensional data such as user's electricity consumption, power load, electricity price fluctuations, climate data, etc. can be collected in real time. The multi-dimensional feature extraction module combines data cleaning and normalization processing methods to ensure that the acquired data is accurate and comprehensive, effectively solving the problem of one-sided and inaccurate traditional user data.

[0254] 2. Significant improvement in the accuracy of user behavior pattern recognition

[0255] Through the deep learning and clustering algorithms in the behavior pattern recognition module, users' electricity usage patterns and habits can be effectively identified to form a detailed user behavior portrait. Through the in-depth mining of user behavior by these algorithms, the system significantly improves the accuracy of user electricity usage behavior pattern recognition and provides more reliable electricity usage behavior analysis for power supply companies and users.

[0256] 3. Improved load forecasting accuracy and optimized demand trend analysis

[0257] The demand forecasting and trend analysis module in the system can accurately predict users' future electricity demand through time series analysis and integrated learning forecasting algorithms. Relevant experimental data show that the load forecasting accuracy of the system is higher than that of traditional forecasting methods, which effectively improves the allocation efficiency of power resources and reduces the risk of unbalanced load in the power grid.

[0258] 4. Personalized service recommendations enhance user power management experience

[0259] Based on user behavior profiles and demand forecast results, the personalized service recommendation module can provide users with tailored electricity usage suggestions. User feedback shows that these suggestions help optimize electricity usage strategies and reduce electricity costs, while helping power supply companies to reasonably adjust electricity price packages and achieve a dynamic balance between user electricity consumption and grid load.

[0260] 5. Data storage and security are significantly enhanced

[0261] The data storage and security module ensures the efficient storage and security of user data through distributed database and encrypted storage technology. Experimental verification shows that the system is superior to traditional methods in data access speed and storage security, and is particularly suitable for the high-frequency and high-density data management needs of smart grids, ensuring user privacy protection and data stability.

[0262] 6. Improve the economic benefits of power supply companies and users

[0263] Through the system's demand forecasting and energy-saving optimization suggestions, users can obtain adaptive energy-saving measures and electricity price packages. Application cases show that the electricity efficiency of enterprises and users has been improved, operating costs and power resource waste have been effectively reduced, bringing economic benefits to power supply companies and helping users save energy costs.

[0264] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0265] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A user behavior data mining and prediction analysis system in a smart grid environment, characterized in that: include: The data collection module is connected with the multi-dimensional feature extraction module and the personalized service recommendation module, and is used to collect user electricity consumption data in real time through smart meters and sensor networks, including electricity consumption, power load, electricity price fluctuations, climate data, and user behavior preferences; The multi-dimensional feature extraction module is connected with the data acquisition module, the behavior pattern recognition module, and the data storage and security module, and is used to extract feature data that can characterize the user's electricity consumption behavior from different dimensions, covering time, space, behavior preferences, and environmental factors, by using data cleaning, preprocessing, and normalization methods; The behavior pattern recognition module is connected with the multi-dimensional feature extraction module and the demand forecasting and trend analysis module to use deep learning and clustering algorithms to identify the patterns and habits of users' electricity consumption behaviors and form user behavior portraits; The demand forecasting and trend analysis module is connected with the behavior pattern recognition module, the personalized service recommendation module, and the data storage and security module to forecast electricity demand through the time series analysis model and the prediction algorithm based on ensemble learning, and output personalized load forecasting results; The personalized service recommendation module is connected to the demand forecasting and trend analysis module to generate personalized electricity usage suggestions for different users based on user behavior profiles and demand forecasts. Power supply companies can provide adaptive electricity price packages and energy-saving measures accordingly. The data storage and security module is connected with the data acquisition module, the multi-dimensional feature extraction module, and the demand forecasting and trend analysis module, and is used to use distributed database and encryption storage technology to ensure efficient storage, query and security of data.

2. The user behavior data mining and prediction analysis system according to claim 1, characterized in that: The demand forecasting and trend analysis module uses a time series forecasting model to forecast the user's future electricity demand and establishes a personalized load forecasting model for the user through an integrated learning algorithm, where the forecast calculation formula for electricity demand is: P(t)=α·f1(t)+β·f2(t)+γ·f3(t) in: P(t) is the predicted electricity demand at time t; f1(t), f2(t), and f3(t) are functions of historical electricity consumption, climate factors, and electricity price fluctuations, respectively; α, β, γ are weight coefficients obtained through training of the ensemble learning algorithm, satisfying α+β+γ=1.

3. The user behavior data mining and prediction analysis system in a smart grid environment as claimed in claim 1, characterized in that: The data acquisition module implementation method: Equipment installation: Install smart meters and various sensors at the user end, including current and voltage sensors, temperature and humidity sensors; Real-time data collection: Smart meters are used to record users’ electricity consumption, voltage, and power load data in real time, and sensors are used to collect environmental and climate data; Utilize IoT technology through wireless transmission protocols including ZigBee and Wi-Fi; Send the collected data to the centralized data collection platform; Multi-source data integration: In the integration platform, the received data is merged and timestamped to generate a data set that can be analyzed.

4. The user behavior data mining and prediction analysis system in a smart grid environment as claimed in claim 1, characterized in that: The multi-dimensional feature extraction module implementation method: Data cleaning and denoising: outliers in the data are processed through outlier processing and interpolation methods to remove invalid or distorted data; Feature normalization: Normalize data from different sources, normalize power data and environmental data to the same numerical range, so that they are consistent in subsequent analysis; Feature engineering: Combine domain knowledge and statistical analysis to generate feature vectors, including peak and valley electricity consumption characteristics, and the relationship between temperature, humidity and electricity consumption; use PCA (principal component analysis) or t-SNE (t-distributed stochastic neighbor embedding) dimensionality reduction methods to further simplify the data, reducing the data dimension while retaining key features.

5. The user behavior data mining and prediction analysis system in a smart grid environment as claimed in claim 1, characterized in that: The behavior pattern recognition module implementation method: Data input: The feature data generated by the multi-dimensional feature extraction module is used as input; Cluster analysis: Use K-means or DBSCAN clustering algorithms to group user data and identify user groups with similar electricity usage patterns, such as those with high peak electricity usage and large seasonal electricity usage variations. Deep learning model: Use LSTM long short-term memory network or convolutional neural network CNN to perform time series analysis on the user's historical electricity consumption data, thereby capturing the user's long-term and short-term electricity consumption behavior; Behavioral profile generation: The model outputs each user's behavioral profile, including electricity usage patterns, electricity usage preferences, and time period preferences, for subsequent demand forecasting and personalized service use.

6. The user behavior data mining and prediction analysis system in a smart grid environment as claimed in claim 1, characterized in that: The demand forecasting and trend analysis module implementation method: Time Series Forecasting Models: A time series data set is constructed based on historical electricity consumption data, and the ARIMA model is used to predict the short-term electricity demand of a single user or a single cell; For long-term data, the LSTM model is used to forecast user electricity demand on a daily, weekly or monthly basis to take into account the changing trends in the long-term time series; Ensemble learning algorithm: Combining random forest and XGBoost algorithms, the prediction results of multiple models are integrated and optimized to ensure the stability and prediction accuracy of the model in complex scenarios; The integrated model can accurately predict future electricity demand under various characteristic conditions, enabling the system to respond in real time under different weather and user behavior changes; Trend analysis: Conduct data trend analysis on the forecast results to identify changes in users' electricity demand, including seasonal fluctuations and temporary peak electricity demand.

7. The user behavior data mining and prediction analysis system in a smart grid environment as claimed in claim 1, characterized in that: The personalized service recommendation module implementation method: Recommendation algorithm: Use collaborative filtering recommendation algorithm, combined with user portrait data and demand prediction results, to push personalized recommendations to different users; Content-based recommendations: Analyze users’ electricity usage habits and behavior profiles, and recommend the best electricity packages and time strategies; Recommendation based on collaborative filtering: Provide group recommendations for users with similar electricity usage patterns; Personalized suggestion generation: By analyzing each user's behavior pattern and load forecast results, different energy-saving plans are pushed to users, including electricity usage suggestions for specific time periods and usage suggestions during peak electricity price periods; User feedback collection: The system supports collecting user feedback and behavior change data, continuously optimizing recommendation algorithms, and improving user experience; Data storage and security module implementation method: Distributed database: Use distributed database to store massive user data to ensure efficient reading, writing and querying of data; distributed storage can process large amounts of real-time data and support horizontal expansion; Encrypted storage: During data storage and transmission, AES and SSL / TLS protocols are used to encrypt user data to ensure user privacy and security; Access control: Through the role-based access control mechanism, ensure that users with different permissions can only access data with corresponding permissions, thereby avoiding the leakage of user data.

8. A method for user behavior data mining and prediction analysis in a smart grid environment implementing the user behavior data mining and prediction analysis system in a smart grid environment as claimed in any one of claims 1 to 6, characterized in that: The user behavior data mining and prediction analysis method in the smart grid environment includes: Step 1: Use smart meters and sensor networks through the data collection module to collect user electricity consumption data in real time, including electricity consumption, power load, electricity price fluctuations, climate data, and user behavior preferences; Step 2: Using data cleaning, preprocessing and normalization methods through the multidimensional feature extraction module, feature data that can characterize the user's electricity consumption behavior is extracted from different dimensions, covering time, space, behavior preferences and environmental factors; Step 3: Use deep learning and clustering algorithms through the behavior pattern recognition module to identify the patterns and habits of the user's electricity usage behavior and form a user behavior profile; Step 4: Use the demand forecasting and trend analysis module to forecast electricity demand using the time series analysis model and the prediction algorithm based on ensemble learning, and output personalized load forecast results; Step 5: Generate personalized electricity usage suggestions for different users based on user behavior profiles and demand forecasts through the personalized service recommendation module, and power supply companies can provide adaptive electricity price packages and energy-saving measures accordingly; Step 6: Use distributed database and encrypted storage technology through the data storage and security module to ensure efficient storage, query and security of data.

9. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the steps of the user behavior data mining and predictive analysis method in a smart grid environment as claimed in claim 8.

10. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the user behavior data mining and predictive analysis system in the smart grid environment as described in any one of claims 1-7.

Images