Multi-factor data prediction method and system, readable storage medium and computer

Abstract

The invention provides a multi-factor data prediction method and system, a readable storage medium and a computer. The method comprises the steps of performing standardization processing on historical monitoring data and external factor data of each monitoring device to obtain standard historical monitoring data and standard external factor data; associating the two pieces of data and inputting the data into a comprehensive prediction model for training to obtain a comprehensive prediction training model; performing data analysis on the real-time data of each monitoring device by adopting a stream processing algorithm, and performing model optimization on the comprehensive prediction training model according to a data analysis result to obtain a comprehensive prediction optimization model; and inputting the to-be-processed scene information into the comprehensive prediction optimization model to obtain a data prediction result. According to the method, the monitoring data and the external factor data are combined, and the seasonal and periodic modes in the data are identified and utilized to construct the multi-factor comprehensive prediction model, so that the prediction accuracy is improved, and the model has obvious advantages in coping with complex and changeable monitoring scenes.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a multi-factor data prediction method, system, readable storage medium and computer. Background Art

[0002] With the development of the Internet of Things and big data technology, all industries can collect a large amount of monitoring data through sensors, monitoring equipment and other means. These data include temperature and humidity, energy consumption, human flow, equipment operation status, etc. They are often closely related to time and may be affected by external factors such as seasons, festivals, and weather. However, most existing technologies are limited to static chart display of data, or short-term prediction based on simple statistical methods such as moving averages and weighted averages. Although these methods can process real-time data and provide short-term predictions to a certain extent, they have many shortcomings when processing long-term, complex, and multi-dimensional data.

[0003] Currently, common data analysis methods rely on existing data sets, mainly including: Data trend analysis: Display the trend of historical data by drawing charts, etc. This method can intuitively display the fluctuation of data, but it cannot deeply explore the rules and cannot effectively predict future data.

[0004] Short-term forecasting: Many industries use short-term forecasting models based on existing data, such as moving average method, exponential smoothing method, etc. These methods calculate short-term data trends (such as data from the last few days or hours) to make forecasts for a period of time in the future. However, these models usually assume that the data trend remains unchanged and ignore long-term factors such as seasonal changes and holiday effects. Therefore, the forecast results can only be applied to short-term scenarios and cannot cope with the impact of external changes.

[0005] There are some obvious limitations in the current existing technologies, especially in the prediction capabilities in long-term and complex scenarios. The main problems are as follows: Unable to summarize the laws of multi-dimensional data: Most existing methods can only process historical data of a single dimension, and lack the ability to summarize the laws of multi-dimensional data (such as time, holidays, seasonal changes, etc.). This makes the model unable to fully utilize different types of external factors (such as holidays, seasonal changes, climate change, etc.), thus affecting the accuracy and stability of the prediction.

[0006] Limitations of short-term forecasting: Most existing methods, such as moving average and exponential smoothing, are mainly aimed at short-term trend forecasting and can usually only effectively predict data changes in the next few days or hours. These methods cannot effectively predict long-term trend changes, such as seasonal fluctuations, cyclical fluctuations, etc., resulting in poor results when long-term planning or cross-seasonal forecasting is required.

[0007] Insufficient processing of time-related data: For example, for data with obvious time correlation (such as temperature, humidity, and pedestrian flow), existing methods find it difficult to accurately identify and utilize the seasonal and cyclical changes in these data, resulting in these data being unable to fully reflect the real trend of change, which in turn affects the accuracy of the prediction results. Summary of the invention Based on this, the purpose of the present invention is to provide a multi-factor data prediction method, system, readable storage medium and computer to solve the deficiencies in the above-mentioned technology.

[0008] The present invention proposes a multi-factor data prediction method, comprising: Acquire historical monitoring data and external factor data of each monitoring device, and perform standardization processing on the historical monitoring data and the external factor data to obtain corresponding standard historical monitoring data and standard external factor data; Constructing a comprehensive prediction model, and correlating the standard historical monitoring data with the standard external factor data, and inputting the correlated data into the comprehensive prediction model for training to obtain a comprehensive prediction training model; Collecting real-time monitoring data and real-time external factor data of each monitoring device in real time, and using a stream processing algorithm to perform data analysis on the real-time monitoring data and the real-time external factor data, and optimizing the comprehensive prediction training model based on the data analysis results to obtain a comprehensive prediction optimization model; The scene information to be processed is obtained, and the scene information to be processed is input into the comprehensive prediction optimization model to obtain the corresponding data prediction result.

[0009] Further, the steps of obtaining historical monitoring data and external factor data of each monitoring device, and standardizing the historical monitoring data and the external factor data to obtain corresponding standard historical monitoring data and standard external factor data include: Obtain historical monitoring data of each monitoring device, and obtain external factor data through data structure and network resources, wherein the external factor data includes holiday information, weather data, and temperature and humidity change data; The external factor data is cleaned and deduplicated, and the processed external factor data is mapped and converted with the historical monitoring data to obtain corresponding standard historical monitoring data and standard external factor data.

[0010] Further, the steps of constructing a comprehensive prediction model, associating the standard historical monitoring data with the standard external factor data, and inputting the associated data into the comprehensive prediction model for training to obtain a comprehensive prediction training model include: Constructing a comprehensive prediction model, and correlating the standard historical monitoring data with the standard external factor data to obtain corresponding correlation data; Optimizing a preset seasonal data prediction algorithm, and integrating the optimized seasonal data prediction algorithm with time series analysis and machine learning algorithms in the comprehensive prediction model to achieve trend prediction and anomaly detection of the comprehensive prediction model in a complex environment; The associated data is input into the comprehensive prediction model for training, and the accuracy of the model is improved through cross-validation and parameter optimization to obtain a comprehensive prediction training model.

[0011] Furthermore, the step of obtaining the scene information to be processed and inputting the scene information to be processed into the comprehensive prediction optimization model to obtain the corresponding data prediction result includes: Obtaining scene information to be processed, and parsing monitoring requirements in the scene information to be processed; The data collection frequency, the priority of external factor data and the model parameters of the comprehensive prediction optimization model are adjusted according to the monitoring requirements, and data prediction is performed on the scene information to be processed to obtain corresponding data prediction results.

[0012] The present invention also proposes a multi-factor data prediction system, comprising: A standardization processing module, used for acquiring historical monitoring data and external factor data of each monitoring device, and performing standardization processing on the historical monitoring data and the external factor data to obtain corresponding standard historical monitoring data and standard external factor data; A model training module is used to construct a comprehensive prediction model, and associate the standard historical monitoring data with the standard external factor data, and input the associated data into the comprehensive prediction model for training to obtain a comprehensive prediction training model; A data analysis module, which is used to collect the real-time monitoring data and real-time external factor data of each of the monitoring devices in real time, and use a stream processing algorithm to perform data analysis on the real-time monitoring data and the real-time external factor data, and use the data analysis results to optimize the comprehensive prediction training model to obtain a comprehensive prediction optimization model; The data prediction module is used to obtain the scene information to be processed and input the scene information to be processed into the comprehensive prediction optimization model to obtain the corresponding data prediction result.

[0013] Furthermore, the standardization processing module includes: A data acquisition unit, used to acquire historical monitoring data of each monitoring device, and acquire external factor data through data structure and network resources, wherein the external factor data includes holiday information, weather data, and temperature and humidity change data; The data processing unit is used to perform data cleaning and data deduplication on the external factor data, and perform data mapping and data conversion on the processed external factor data and the historical monitoring data to obtain corresponding standard historical monitoring data and standard external factor data.

[0014] Furthermore, the model training module includes: A data association unit, used to construct a comprehensive prediction model and associate the standard historical monitoring data with the standard external factor data to obtain corresponding associated data; An algorithm optimization unit, used to optimize a preset seasonal data prediction algorithm, and integrate the optimized seasonal data prediction algorithm with a time series analysis and a machine learning algorithm in the comprehensive prediction model to achieve trend prediction and anomaly detection of the comprehensive prediction model in a complex environment; The model training unit is used to input the associated data into the comprehensive prediction model for training, and improve the accuracy of the model through cross-validation and parameter optimization to obtain a comprehensive prediction training model.

[0015] Furthermore, the data prediction module includes: A data analysis unit, used to obtain the scene information to be processed and analyze the monitoring requirements in the scene information to be processed; The data prediction unit is used to adjust the data collection frequency, the priority of external factor data and the model parameters of the comprehensive prediction optimization model according to the monitoring requirements, and perform data prediction on the scene information to be processed to obtain corresponding data prediction results.

[0016] The present invention also provides a readable storage medium on which a computer program is stored. When the program is executed by a processor, the multi-factor data prediction method described above is implemented.

[0017] The present invention also provides a computer, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned multi-factor data prediction method when executing the computer program.

[0018] The multi-factor data prediction method, system, readable storage medium and computer in the present invention improve the accuracy of the prediction by combining monitoring data and external factor data, and identifying and utilizing seasonal and cyclical patterns in the data to construct a multi-factor comprehensive prediction model, so that the model has obvious advantages in dealing with complex and changeable monitoring scenarios; the seasonal data prediction algorithm is optimized so that the model can better capture the periodicity and trend changes in the data, can automatically detect long-term trends and short-term fluctuations in the data, and can effectively perform accurate predictions even in the presence of noise and outliers in the data, and can more comprehensively reflect the laws of data changes in reality. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flow chart of a multi-factor data prediction method in a first embodiment of the present invention; Figure 2 for Figure 1 Detailed flow chart of step S101; Figure 3 for Figure 1 Detailed flow chart of step S102; Figure 4 for Figure 1 Detailed flow chart of step S104; Figure 5 is a structural block diagram of a multi-factor data prediction system in a second embodiment of the present invention; Figure 6 FIG. 4 is a structural block diagram of a computer in a third embodiment of the present invention.

[0020] The following specific implementation manner will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0021] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0023] Embodiment 1 See also Figure 1, which shows a multi-factor data prediction method in a first embodiment of the present invention, and the method specifically includes steps S101 to S104: S101, acquiring historical monitoring data and external factor data of each monitoring device, and performing standardization processing on the historical monitoring data and the external factor data to obtain corresponding standard historical monitoring data and standard external factor data; For further information, see Figure 2 , the step S101 specifically includes steps S1011~S1012: S1011, obtaining historical monitoring data of each monitoring device, and obtaining external factor data through data structure and network resources, wherein the external factor data includes holiday information, weather data, and temperature and humidity change data; S1012, performing data cleaning and data deduplication on the external factor data, and performing data mapping and data conversion on the processed external factor data and the historical monitoring data to obtain corresponding standard historical monitoring data and standard external factor data.

[0024] In the specific implementation, by connecting with various detection equipment (such as temperature and humidity sensors, weather stations, and flow monitors, etc.), relevant data, including temperature and humidity, air quality, flow, equipment status, etc., are collected in real time. Corresponding data is collected according to different scenarios to ensure that the types of collected data are complete and the frequency meets the requirements, covering all key parameters of the target scenario; Furthermore, data on external influencing factors are regularly obtained through data interfaces and network resources, including holiday information, weather forecasts, temperature and humidity changes, etc., and pre-processing steps such as data cleaning and deduplication are used to ensure data integrity and consistency.

[0025] Standardize data from different sources to ensure effective integration during analysis and modeling. Specifically, use data mapping and transformation to achieve consistency in multi-dimensional data.

[0026] S102, constructing a comprehensive prediction model, and correlating the standard historical monitoring data with the standard external factor data, and inputting the correlated data into the comprehensive prediction model for training to obtain a comprehensive prediction training model; For further information, see Figure 3 , the step S102 specifically includes steps S1021 to S1023: S1021, constructing a comprehensive prediction model, and correlating the standard historical monitoring data with the standard external factor data to obtain corresponding correlation data; S1022, optimizing a preset seasonal data prediction algorithm, and integrating the optimized seasonal data prediction algorithm with a time series analysis and a machine learning algorithm in the comprehensive prediction model, so as to realize trend prediction and anomaly detection of the comprehensive prediction model in a complex environment; S1023, inputting the associated data into the comprehensive prediction model for training, and improving the accuracy of the model through cross-validation and parameter optimization to obtain a comprehensive prediction training model.

[0027] In the specific implementation, a multi-factor comprehensive prediction model is constructed to associate the monitoring data with external factors (such as time, holidays, weather, temperature and humidity, etc.). First, the monitoring data and external factors (such as time, holidays, weather, etc.) are effectively preprocessed and feature extracted. Time features (such as seasons, months, weeks, etc.) and external factors (such as weather, holidays, etc.) need to be aligned with monitoring data (such as temperature and humidity, flow, etc.) to form a unified input matrix. The key in feature engineering is to extract seasonal and periodic information from timestamps, such as adding seasons, holidays, etc. as features to the model.

[0028] ARIMA or SARIMA is used to handle trends and seasonal changes, LSTM or GRU is used to capture complex long-term dependencies and cyclical fluctuations, and random forest and XGBoost are used to handle multi-dimensional external factors with an integrated learning approach.

[0029] The core of the model is the ability to identify and utilize seasonal, cyclical and trend changes in the data.

[0030] Furthermore, an improved seasonal data prediction algorithm is used, in which the improved algorithm combines time series analysis and deep learning models, captures long-term trends, short-term fluctuations and seasonal patterns in the data through multi-factor feature input and adaptive mechanisms, and incorporates external influencing factors (such as holidays, weather, etc.) for prediction. The algorithm uses a weighted integration model to combine the linear prediction of traditional statistical models (such as SARIMA) with the nonlinear prediction results of deep learning models (such as LSTM) to optimize the overall prediction accuracy. The formula of the improved seasonal data prediction algorithm is as follows: ; In the formula, Indicates time The predicted value of represents the linear trend and seasonal forecast based on the SARIMA model, Represents nonlinear trend prediction based on LSTM model, combined with time series data and external factors data , It represents the correction item of the forecast value caused by external impression data (such as holidays and weather). , , Respectively represent the weight parameters of the model, satisfying , and make adjustments through checking and verifying historical data. By integrating time series analysis and machine learning algorithms (LSTM, ARIMA, etc.) into the model, trend prediction and anomaly detection in complex environments are achieved, the model is trained based on historical data, and the accuracy of the model is improved through cross-validation, parameter optimization, hyperparameter adjustment, etc. Ensure that the model can flexibly adapt to changing external factors and strike a balance between prediction accuracy and real-time performance.

[0031] Specifically, cross-validation is used to determine the generalization ability of a model. It divides the dataset into multiple subsets (usually 5 or 10 folds), trains the model on each fold and validates it on the remaining folds, and finally averages the validation results of each fold to obtain the final performance of the model. The specific steps are as follows: 1. Split data: Divide the data set into k subsets in proportion, usually k=5 or k=10.

[0032] 2. Training and validation: For each subset, use the remaining k−1 subsets to train the model and use the current subset for validation.

[0033] 3. Performance evaluation: Record the error of each verification and calculate the average error of all folds.

[0034] 4. Model selection: Select the optimal model based on the cross-validation results.

[0035] Assuming K-fold cross validation is used, the cross validation error can be expressed as:

[0036] In the formula, Indicates The error on the fold (such as mean square error MSE) can be effectively In this embodiment, a multi-factor comprehensive prediction model is constructed by combining the monitoring data with external factors such as holidays, seasons, temperature and humidity, and weather. The model can identify and utilize seasonal and cyclical changes in the data to improve the accuracy and adaptability of the prediction.

[0037] Adopt the time series decomposition method: decompose the time series data into four parts: trend, seasonality, cyclicity, and noise, to help the model focus on learning the seasonal and cyclical parts.

[0038] Suppose you are predicting the average monthly temperature data for a certain area. The data may show regular changes throughout the year, which is seasonal change. Through time series decomposition methods, the model can separate seasonal patterns from the overall data, such as the trend of rising temperatures in summer and falling temperatures in winter every year. In this way, the model can accurately identify this seasonal change during training.

[0039] The analysis process includes three parts: 1. Trend component: Captures the long-term trend of data changes (for example, the temperature gradually rises year by year).

[0040] 2. Seasonality: Captures periodic fluctuations in the data (e.g., high temperatures in summer and low temperatures in winter).

[0041] 3. Noise: Some random fluctuations may exist, which are usually ignored.

[0042] The time series decomposition adopts an additive model:

[0043] In the formula, Represents the original data, Represents the trend component, represents the seasonal component, Indicates noise.

[0044] S103, collecting real-time monitoring data and real-time external factor data of each monitoring device in real time, and using a stream processing algorithm to perform data analysis on the real-time monitoring data and the real-time external factor data, and optimizing the comprehensive prediction training model based on the data analysis results to obtain a comprehensive prediction optimization model; In the specific implementation, stream processing technology (such as Apache Kafka algorithm or Kafka Streams algorithm) is used to realize the rapid collection and analysis of real-time data of each monitoring device, and process data at the moment of data generation to meet the high real-time requirements in the monitoring scenario. When an abnormal trend or fluctuation of a certain data is detected, the prediction strategy will be automatically adjusted and the results will be fed back to the user.

[0045] Furthermore, the above data analysis results are used to optimize the comprehensive forecast training model to obtain a comprehensive forecast optimization model. The sliding window and incremental learning methods are combined to enable the model to dynamically adapt to the changing seasonal and cyclical data, ensuring the forecast accuracy and improving the learning efficiency of the model. The model learning process can be divided into the following steps: 1. Introduction of sliding window mechanism First, the model uses a sliding window method to dynamically select training data. The sliding window slices the time series data by setting a fixed time window (for example, the data of the past 30 days). Whenever a new data point arrives, the window will slide forward and replace the earliest historical data (taking the data of the new day and the previous 29 days), always maintaining the timeliness of the data in the window. This mechanism ensures that the model can reflect the latest data changes in a timely manner, especially seasonal fluctuations and cyclical changes.

[0046] 2. Gradual Optimization of Incremental Learning Within each window, an incremental learning method is used to update the model. Whenever new data is introduced, the incremental learning algorithm fine-tunes the existing model based on the new data instead of retraining the entire model. In this way, the model can effectively utilize historical knowledge while continuously absorbing and adapting to new data, thereby maintaining high prediction accuracy. The core of incremental learning is to adjust the weights of the model through parameter updates so that it can continuously learn new data patterns. For example, in the temperature prediction model, with the arrival of each batch of new temperature data, the model adjusts the weights of temperature-related features to better adapt to seasonal changes.

[0047] 3. Dynamically adjust model weights and parameters Combining sliding windows and incremental learning, the model can flexibly adjust weights and parameters to cope with seasonal, cyclical and trend changes in the data. When new data points enter the window and incremental learning is performed, the model's weights are dynamically adjusted according to the following methods: Seasonal changes: Through incremental learning, the model can identify and strengthen the weights of features related to seasonal changes, such as the changing patterns of winter and summer temperatures. If new data indicates that summer temperatures are abnormally high, the model will increase the weight of the temperature feature, making the temperature more influential on the prediction results.

[0048] Cyclical changes: Cyclical changes are also gradually captured through incremental learning. As data is introduced each year, the model gradually adjusts the weights of cyclical features so that the prediction results can adapt to cyclical fluctuations.

[0049] Trend changes: Trend data changes (such as increasing temperatures or increasing electricity demand) are adapted through parameter updates in incremental learning. Over time, the model continuously adjusts weights to strengthen long-term trends.

[0050] Understandably, the model dynamically adjusts weights and parameters according to data changes in actual scenarios to make predictions more accurate.

[0051] S104, obtaining scene information to be processed, and inputting the scene information to be processed into the comprehensive prediction optimization model to obtain corresponding data prediction results.

[0052] For further information, see Figure 4 , the step S104 specifically includes steps S1041-S1042: S1041, obtaining scene information to be processed, and parsing monitoring requirements in the scene information to be processed; S1042, adjusting the data collection frequency, the priority of the external factor data and the model parameters of the comprehensive prediction optimization model according to the monitoring requirements, and performing data prediction on the scene information to be processed to obtain corresponding data prediction results.

[0053] In the specific implementation, obtain the scene information to be processed: In the data prediction process, firstly, it is necessary to collect the data of the scene to be processed through the relevant system (such as sensors, monitoring equipment or data acquisition module). These data may include external factors such as time, date, weather conditions, temperature and humidity, holiday information, etc., and also include the monitored target data (such as temperature, energy consumption, vibration, etc.). These data reflect the actual situation that is currently or about to occur.

[0054] Specifically, the data is preprocessed. The acquired raw data usually needs to be preprocessed, including removing noise, filling missing values, data normalization, etc. This step ensures that the input data reaches the best state before prediction, so that the model can process it effectively.

[0055] Furthermore, the scenario information is input into the comprehensive prediction optimization model, including the following steps: 1. Data input and model interface: Preprocessed scene information (such as time features, weather data, holidays, etc.) will be input into the comprehensive forecasting optimization model. Based on the learning results of historical data, the model can identify the correlation and regularity between different features.

[0056] 2. Model identification and association processing: After the data is input, the model will automatically analyze the characteristics of the input data (such as season, cycle, etc.), combining the sliding window mechanism and incremental learning method to identify seasonal changes, cyclical fluctuations, etc. in the current scenario. This process utilizes the patterns accumulated in historical data and combines the latest data for prediction.

[0057] Furthermore, the processing of the comprehensive prediction optimization model is as follows: 1. Prediction and dynamic adjustment: Based on the input scenario information, the comprehensive prediction optimization model will dynamically adjust the model parameters and weights according to the aforementioned incremental learning mechanism. For example, if the data indicates that the temperature is rising, the model may automatically adjust the weight of the temperature feature to improve the accuracy of the prediction results. As new data is continuously input, the model will update and optimize the prediction results in each sliding window.

[0058] 2. Seasonal and cyclical modeling: The model can automatically identify and apply seasonal changes, cyclical fluctuations, and trend changes. For example, if it is summer, the model will identify the cyclical relationship between high temperatures and demand surges, thereby improving the forecast accuracy of electricity demand.

[0059] Furthermore, the corresponding data prediction results are obtained. Through the processing and analysis of the scene data by the model, the corresponding prediction results will be output. These results may be the future value of a specific indicator (such as the temperature in the next few days), and detailed prediction data will be given according to the needs.

[0060] Specifically, real-time feedback and model updates: With the continuous input of new scene information and data, the comprehensive prediction optimization model will be continuously adjusted and optimized. This combination of incremental learning and sliding windows ensures the adaptability and accuracy of the model under different environmental changes.

[0061] In this embodiment, a system based on a C / S architecture is constructed, in which the server is responsible for real-time processing, analysis and storage of data, and the client is responsible for display and interaction, thereby improving the response speed of the system and ensuring the stability of data processing.

[0062] It is understandable that in some optional embodiments, for scenarios with low real-time requirements, the architecture can also be a B / S architecture, which displays data through web pages and is suitable for enterprises with high requirements on system development costs.

[0063] Furthermore, an intuitive graphical user interface (GUI) is designed on the client side, including trend charts, bar charts, pie charts and other chart forms to visually display monitoring data and prediction results. Users can view the changes in various types of data and the prediction analysis results of the model in real time through the interface. In this embodiment, multiple alarm rules are set, such as abnormal data fluctuations, exceeding thresholds, etc. When an anomaly is detected, the system will promptly send an alarm notification to the user to help the user respond quickly. For example, when the data exceeds a preset safety range or a possible abnormal situation is predicted, the system will automatically send a notification to the user, and the user can receive real-time reminders through mobile applications or emails so that timely countermeasures can be taken.

[0064] Specifically, independent configuration modules are designed for different application scenarios (such as traffic flow monitoring, environmental quality monitoring, production management, etc.). Users can flexibly adjust the frequency of data collection, the priority of external factors, model parameters, etc. according to specific monitoring needs to meet the different needs of various industries. In this embodiment, a modular design is adopted to enable the prediction model and data processing module to be independently developed and expanded. Users can choose different analysis modules to combine and replace according to actual needs, further improving the versatility and scalability of the system.

[0065] This embodiment can better capture the trends and periodic patterns in time series data by designing an improved seasonal data prediction algorithm. The algorithm performs well in dealing with complex and changeable monitoring scenarios, effectively improving the efficiency and accuracy of predictions, and uniformly collects and integrates multi-dimensional data from different monitoring devices (such as temperature and humidity sensors, weather station data, flow monitors, etc.) to establish a correlation analysis model. Through the integration of cross-data sources, the system can more comprehensively understand the complex relationships in the monitoring scenarios, identify potential trends and anomalies, and effectively improve the analysis depth and accuracy of monitoring data.

[0066] In this embodiment, the model of the present application has the ability of self-learning and dynamic adjustment, and can automatically optimize parameter configuration according to changes in actual monitoring data. This adaptability reduces the need for human intervention and enables the system to maintain high performance for a long time. In addition, multi-scenario adaptability and flexible configuration modules enable the system to be widely used in multiple fields such as transportation, environment, and production. Users can freely select and adjust monitoring parameters and external factors according to their needs, enhancing the versatility of the system.

[0067] Furthermore, through the real-time data display system based on C / S architecture, this application can provide users with an intuitive graphical user interface, and various monitoring data and prediction results are clearly presented in the form of charts, simplifying the data analysis and operation process. Compared with traditional command line operations or complex data analysis tools, users can use it conveniently without professional technical background, which improves the user experience and operation convenience of the system.

[0068] In summary, the multi-factor data prediction method in the above-mentioned embodiment of the present invention improves the accuracy of the prediction by combining the monitoring data and the external factor data, and identifying and utilizing the seasonal and cyclical patterns in the data to construct a multi-factor comprehensive prediction model, so that the model has obvious advantages in dealing with complex and changeable monitoring scenarios; the seasonal data prediction algorithm is optimized so that the model can better capture the periodicity and trend changes in the data, and can automatically detect the long-term trends and short-term fluctuations in the data, and even in the presence of noise and outliers in the data, it can effectively perform accurate predictions, and can more comprehensively reflect the laws of data changes in reality.

[0069] Embodiment 2 Another aspect of the present invention is to provide a multi-factor data prediction system. Figure 5 , which is a multi-factor data prediction system in a second embodiment of the present invention, and the system comprises: The standardization processing module 11 is used to obtain the historical monitoring data and external factor data of each monitoring device, and perform standardization processing on the historical monitoring data and the external factor data to obtain corresponding standard historical monitoring data and standard external factor data; Furthermore, the standardization processing module 11 includes: A data acquisition unit, used to acquire historical monitoring data of each monitoring device, and acquire external factor data through data structure and network resources, wherein the external factor data includes holiday information, weather data, and temperature and humidity change data; The data processing unit is used to perform data cleaning and data deduplication on the external factor data, and perform data mapping and data conversion on the processed external factor data and the historical monitoring data to obtain corresponding standard historical monitoring data and standard external factor data.

[0070] The model training module 12 is used to construct a comprehensive prediction model, and associate the standard historical monitoring data with the standard external factor data, and input the associated data into the comprehensive prediction model for training to obtain a comprehensive prediction training model; Furthermore, the model training module 12 includes: A data association unit, used to construct a comprehensive prediction model and associate the standard historical monitoring data with the standard external factor data to obtain corresponding associated data; An algorithm optimization unit, used to optimize a preset seasonal data prediction algorithm, and integrate the optimized seasonal data prediction algorithm with a time series analysis and a machine learning algorithm in the comprehensive prediction model to achieve trend prediction and anomaly detection of the comprehensive prediction model in a complex environment; The model training unit is used to input the associated data into the comprehensive prediction model for training, and improve the accuracy of the model through cross-validation and parameter optimization to obtain a comprehensive prediction training model.

[0071] The data analysis module 13 is used to collect the real-time monitoring data and real-time external factor data of each monitoring device in real time, and use the stream processing algorithm to perform data analysis on the real-time monitoring data and the real-time external factor data, and optimize the comprehensive prediction training model with the data analysis results to obtain a comprehensive prediction optimization model; The data prediction module 14 is used to obtain the scene information to be processed and input the scene information to be processed into the comprehensive prediction optimization model to obtain the corresponding data prediction result.

[0072] Furthermore, the data prediction module 14 includes: A data analysis unit, used to obtain the scene information to be processed and analyze the monitoring requirements in the scene information to be processed; The data prediction unit is used to adjust the data collection frequency, the priority of external factor data and the model parameters of the comprehensive prediction optimization model according to the monitoring requirements, and perform data prediction on the scene information to be processed to obtain corresponding data prediction results.

[0073] The functions or operation steps implemented when the above modules and units are executed are generally the same as those in the above method embodiments, and will not be repeated here.

[0074] The multi-factor data prediction system provided in the embodiment of the present invention has the same implementation principle and technical effects as those of the aforementioned method embodiment. For the sake of brief description, for matters not mentioned in the system embodiment, reference may be made to the corresponding contents in the aforementioned method embodiment.

[0075] Embodiment 3 The present invention also provides a computer, see Figure 6 , shown is a computer in the third embodiment of the present invention, including a memory 10, a processor 20, and a computer program 30 stored in the memory 10 and executable on the processor 20. When the processor 20 executes the computer program 30, the above-mentioned multi-factor data prediction method is implemented.

[0076] The memory 10 includes at least one type of readable storage medium, which includes a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., an SD or DX memory, etc.), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 10 may be an internal storage unit of a computer, such as a hard disk of the computer. In other embodiments, the memory 10 may also be an external storage device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card, etc. Further, the memory 10 may also include both an internal storage unit of the computer and an external storage device. The memory 10 may be used not only to store application software and various types of data installed in the computer, but also to temporarily store data that has been output or is to be output.

[0077] Among them, in some embodiments, the processor 20 can be an electronic control unit (Electronic Control Unit, abbreviated as ECU, also known as a vehicle computer), a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, a microprocessor or other data processing chip, used to run the program code stored in the memory 10 or process data, such as executing access restriction programs, etc.

[0078] It should be pointed out that Figure 6 The structure shown does not constitute a limitation on the computer. In other embodiments, the computer may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0079] The embodiment of the present invention further provides a readable storage medium having a computer program stored thereon, and when the program is executed by a processor, the multi-factor data prediction method as described above is implemented.

[0080] Those skilled in the art will appreciate that the logic and / or steps represented in the flowchart or otherwise described herein, for example, may be considered as an ordered list of executable instructions for implementing logical functions, and may be specifically implemented in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For purposes of this specification, "computer-readable medium" may be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0081] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0082] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or a combination thereof: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0083] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0084] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A multi-factor data prediction method, characterized in that: include: Acquire historical monitoring data and external factor data of each monitoring device, and perform standardization processing on the historical monitoring data and the external factor data to obtain corresponding standard historical monitoring data and standard external factor data; Constructing a comprehensive prediction model, and correlating the standard historical monitoring data with the standard external factor data, and inputting the correlated data into the comprehensive prediction model for training to obtain a comprehensive prediction training model; Collecting real-time monitoring data and real-time external factor data of each monitoring device in real time, and using a stream processing algorithm to perform data analysis on the real-time monitoring data and the real-time external factor data, and optimizing the comprehensive prediction training model based on the data analysis results to obtain a comprehensive prediction optimization model; The scene information to be processed is obtained, and the scene information to be processed is input into the comprehensive prediction optimization model to obtain the corresponding data prediction result.

2. The multi-factor data prediction method according to claim 1, characterized in that: The steps of obtaining historical monitoring data and external factor data of each monitoring device, and performing standard processing on the historical monitoring data and the external factor data to obtain corresponding standard historical monitoring data and standard external factor data include: Obtain historical monitoring data of each monitoring device, and obtain external factor data through data structure and network resources, wherein the external factor data includes holiday information, weather data, and temperature and humidity change data; The external factor data is cleaned and deduplicated, and the processed external factor data is mapped and converted with the historical monitoring data to obtain corresponding standard historical monitoring data and standard external factor data.

3. The multi-factor data prediction method according to claim 1, characterized in that: The steps of constructing a comprehensive prediction model, associating the standard historical monitoring data with the standard external factor data, and inputting the associated data into the comprehensive prediction model for training to obtain a comprehensive prediction training model include: Constructing a comprehensive prediction model, and correlating the standard historical monitoring data with the standard external factor data to obtain corresponding correlation data; Optimizing a preset seasonal data prediction algorithm, and integrating the optimized seasonal data prediction algorithm with time series analysis and machine learning algorithms in the comprehensive prediction model to achieve trend prediction and anomaly detection of the comprehensive prediction model in a complex environment; The associated data is input into the comprehensive prediction model for training, and the accuracy of the model is improved through cross-validation and parameter optimization to obtain a comprehensive prediction training model.

4. The multi-factor data prediction method according to claim 1, characterized in that: The steps of obtaining the scene information to be processed and inputting the scene information to be processed into the comprehensive prediction optimization model to obtain the corresponding data prediction results include: Obtaining scene information to be processed, and parsing monitoring requirements in the scene information to be processed; The data collection frequency, the priority of external factor data and the model parameters of the comprehensive prediction optimization model are adjusted according to the monitoring requirements, and data prediction is performed on the scene information to be processed to obtain corresponding data prediction results.

5. A multi-factor data prediction system, characterized in that: include: A standardization processing module, used for acquiring historical monitoring data and external factor data of each monitoring device, and performing standardization processing on the historical monitoring data and the external factor data to obtain corresponding standard historical monitoring data and standard external factor data; A model training module is used to construct a comprehensive prediction model, and associate the standard historical monitoring data with the standard external factor data, and input the associated data into the comprehensive prediction model for training to obtain a comprehensive prediction training model; A data analysis module, which is used to collect the real-time monitoring data and real-time external factor data of each of the monitoring devices in real time, and use a stream processing algorithm to perform data analysis on the real-time monitoring data and the real-time external factor data, and use the data analysis results to optimize the comprehensive prediction training model to obtain a comprehensive prediction optimization model; The data prediction module is used to obtain the scene information to be processed and input the scene information to be processed into the comprehensive prediction optimization model to obtain the corresponding data prediction result.

6. The multi-factor data prediction system according to claim 5, characterized in that: The standardization processing module includes: A data acquisition unit, used to acquire historical monitoring data of each monitoring device, and acquire external factor data through data structure and network resources, wherein the external factor data includes holiday information, weather data, and temperature and humidity change data; The data processing unit is used to perform data cleaning and data deduplication on the external factor data, and perform data mapping and data conversion on the processed external factor data and the historical monitoring data to obtain corresponding standard historical monitoring data and standard external factor data.

7. The multi-factor data prediction system according to claim 5, characterized in that: The model training module includes: A data association unit, used to construct a comprehensive prediction model and associate the standard historical monitoring data with the standard external factor data to obtain corresponding associated data; An algorithm optimization unit, used to optimize a preset seasonal data prediction algorithm, and integrate the optimized seasonal data prediction algorithm with a time series analysis and a machine learning algorithm in the comprehensive prediction model to achieve trend prediction and anomaly detection of the comprehensive prediction model in a complex environment; The model training unit is used to input the associated data into the comprehensive prediction model for training, and improve the accuracy of the model through cross-validation and parameter optimization to obtain a comprehensive prediction training model.

8. The multi-factor data prediction system according to claim 5, characterized in that: The data prediction module comprises: A data analysis unit, used to obtain the scene information to be processed and analyze the monitoring requirements in the scene information to be processed; The data prediction unit is used to adjust the data collection frequency, the priority of external factor data and the model parameters of the comprehensive prediction optimization model according to the monitoring requirements, and perform data prediction on the scene information to be processed to obtain corresponding data prediction results.

9. A readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the multi-factor data prediction method according to any one of claims 1 to 4 is implemented.

10. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the multi-factor data prediction method according to any one of claims 1 to 4 is implemented.

Images