Cross-domain big data fusion analysis and intelligent prediction system

Abstract

The invention relates to the technical field of prediction adjustment, in particular to a cross-domain big data fusion analysis and intelligent prediction system, which comprises the following modules: a data acquisition module used for collecting space and time data from multiple surveying and mapping domains; the data preprocessing module is used for cleaning, standardizing and formatting the collected data; the data integration module is used for fusing and unifying data from different surveying and mapping domains and providing a unified data access interface; the intelligent prediction module is used for carrying out trend analysis and prediction on the integrated data by adopting a machine learning algorithm; the monitoring and sensing module collects environmental condition data in real time through an environmental sensor network and combines the environmental condition data with surveying and mapping data; and the adjusting module automatically adjusts a data acquisition strategy and analysis parameters according to the intelligent prediction and monitoring result, and optimizes a data processing flow. According to the method, current and future environmental conditions are reflected more accurately, and the comprehensive prediction method can capture environmental changes in time, so that more accurate support is provided for decisions related to geographic information.

Description

Technical Field

[0001] The present invention relates to the field of prediction and regulation technology, and in particular to a cross-domain big data fusion analysis and intelligent prediction system. Background Art

[0002] In today's data-driven era, big data technology has become the core support for many industries, especially in the field of surveying and mapping, where geographic information systems (GIS), satellite remote sensing, drone measurement and ground station technology are widely used. These technologies can provide a large amount of data on geography, climate and environmental changes, but the challenge that comes with it is how to effectively integrate and analyze surveying and mapping data from different sources and in different formats.

[0003] Traditional data processing systems often face problems such as low data integration efficiency, insufficient prediction accuracy, and slow response to environmental changes. Especially when it is necessary to respond to rapid environmental changes, such as natural disasters or rapidly developing urban areas, existing data processing methods often cannot provide timely and accurate information to support decision-making. In addition, rapid changes in environmental conditions, such as climate fluctuations and surface changes, also require prediction systems to have higher adaptability and flexibility.

[0004] In addition, although there are many attempts to use machine learning and artificial intelligence algorithms for data analysis and prediction in the existing technology, these attempts usually lack effective mechanisms to dynamically adjust data collection and analysis strategies to adapt to environmental changes. This limits the effectiveness of the system in practical applications, especially when dealing with large-scale, cross-domain surveying and mapping data.

[0005] Therefore, there is an urgent need to develop a system that can effectively integrate cross-domain mapping data, achieve high-precision intelligent prediction, and dynamically adjust its data processing strategy according to environmental changes. Summary of the invention

[0006] Based on the above objectives, the present invention provides a cross-domain big data fusion analysis and intelligent prediction system.

[0007] The cross-domain big data fusion analysis and intelligent prediction system includes the following modules: Data acquisition module for collecting spatial and temporal data from multiple mapping domains; Data preprocessing module, used to clean, standardize and format the collected data; Data integration module, used to integrate and unify data from different surveying and mapping domains, and provide a unified data access interface; Intelligent prediction module, which uses machine learning algorithms to perform trend analysis and prediction on integrated data; The monitoring and perception module collects environmental condition data (climate change, surface change) in real time through an environmental sensor network and combines it with surveying and mapping data to enhance the environmental adaptability and accuracy of the prediction; The adjustment module automatically adjusts the data collection strategy and analysis parameters according to the intelligent prediction and monitoring results, and optimizes the data processing process.

[0008] Furthermore, the multiple surveying and mapping domains include geographic information system GIS, satellite remote sensing, drone measurement and ground measurement stations, and the data acquisition module specifically includes: Interface integration mechanism: It has a data interface compatible with multiple surveying and mapping domains and can simultaneously receive multiple data formats, including vector data, raster data, and real-time streaming data; Data filtering and preliminary classification: During the data collection process, implement automated data quality assessment and filtering to eliminate erroneous or incomplete data input, and perform preliminary classification according to data source and type.

[0009] Furthermore, the data integration module includes automatically aligning spatial data from different sources and at different scales, using geocoding and coordinate conversion techniques to ensure that all data are represented in the same spatial reference, and using time synchronization technology to ensure that data from different surveying and mapping domains have consistent time tags to obtain time series data.

[0010] Furthermore, the machine learning algorithm in the intelligent prediction module adopts the ARIMA model (autoregressive integrated moving average model), which specifically includes: Apply difference processing to the time series data after integration by the data integration module to achieve stationarity requirements, especially for those data that show high autocorrelation and seasonal fluctuations. For seasonally varying data, such as satellite data affected by cloud cover in a particular season, seasonal adjustment methods are applied to eliminate periodic effects; Model parameter identification and optimization: Based on the ACF and PACF plots of the integrated data, identify the ARIMA model parameters p, d, q and seasonal parameters P, D, Q, S, use historical data sets to train the ARIMA model, and select the optimal model using the Akaike Information Criterion AIC; The optimized ARIMA model is used to predict future trends in geographic information changes, including ground deformation and vegetation cover changes. The prediction results are input into the adjustment module to automatically adjust the data collection strategy and analysis parameters to optimize the data processing process.

[0011] Furthermore, the difference processing is used to make the non-stationary time series data stationary, as follows: First-order difference: Removes the linear trend of the data and is calculated as: in, is a time series at a point in time The observed value of is the value after the first-order difference. If the data is still not stable after the first-order difference, the second-order difference is performed: ; The seasonal adjustment method is based on seasonal differences and is used to process time series data that show seasonal changes, removing changes within a fixed seasonal cycle and making the series stable on the seasonal cycle. Suppose the seasonal cycle is (For annual periodic data, Take 12), the seasonal difference is expressed as: ,in, is the value after seasonal difference.

[0012] Furthermore, the autocorrelation function ACF shows the correlation between the time series and itself at different lags, which is used to identify the order q of the MA model; the partial autocorrelation function PACF diagram shows the correlation between the time series and itself under the influence of given other lag values, which is used to identify the order p of the AR model; The identification of the non-seasonal parameters p, d, q is as follows: p: Observe the PACF plot and find the lag number with obvious truncation (sudden drop to 0 and close to 0), which indicates the order of the AR part; q: Observe the ACF plot and find the number of lags that are obviously truncated, indicating the order of the MA part; d: The difference order d is determined by the ADF test until the time series becomes stationary; The identification of seasonal parameters P, D, Q, S is as follows: S: seasonal period, determined according to the periodicity of the dataset, such as annually, quarterly, or monthly; P, Q: Same as non-seasonal parameters, determined from seasonal PACF and ACF plots, respectively; D: seasonal difference order, determined by seasonal stationarity test; The ARIMA model uses a historical data set and is trained according to determined model parameters to construct the ARIMA model, including dividing the data into a training set and a test set, applying difference and seasonal difference processing, and ensuring the stationarity of the model input data.

[0013] Furthermore, the Akaike Information Criterion AIC is expressed as: ,in, the number of model parameters, It is the maximum likelihood estimate of the model. During the model training process, the AIC value is calculated for different parameter combinations, and the model with the lowest AIC value is selected as the optimal model, because a lower AIC value indicates a better balance between the model's fitting quality and complexity.

[0014] Furthermore, the ARIMA model is expressed as ARIMA(p, d, q), where p is the order of the autoregressive term, indicating the number of past values ​​used in the model to predict the current value, d is the number of differences, indicating the number of non-seasonal differences required to make the time series stationary, and q is the order of the moving average term, indicating the number of past prediction errors used to predict the current value. The model is expressed as: ,in, It's time The original sequence of is the parameter of the autoregressive term, is the parameter of the moving average term, is the lag operator , is the white noise error term; The seasonal ARIMA is expressed as SARIMA. When the time series data show obvious seasonal fluctuations, SARIMA is used, which is expressed as , where P, D, Q: represent the orders of seasonal autoregressive term, seasonal difference and seasonal moving average term respectively, S represents the seasonal length of the time series, and SARIMA is expressed as: ,in, and are the parameters of the seasonal autoregressive and seasonal moving average terms, is the seasonal lag operator, , for or .

[0015] Furthermore, the monitoring and perception module specifically includes: Environmental data integration: real-time environmental condition data collected from environmental sensor networks are synchronized in time and space with surveying and mapping data to ensure that all data items are aligned before analysis. Environmental condition data, including climate factors (temperature, humidity, rainfall) and surface characteristics (vegetation coverage, terrain changes), are feature extracted and converted into a format that can be used for model input; The environmental characteristic data are fused with the integrated surveying and mapping data to create an extended dataset that contains not only time series data but also environmental variables to provide comprehensive prediction factors. Differentiation and transformation processing are applied to the extended dataset to ensure that the data meets the stationarity requirements of the ARIMA model. ARIMA model enhancement: Adjust the parameters (p, d, q) of the ARIMA model based on the extended dataset, consider the impact of environmental factors on the model dynamics, re-evaluate the seasonal parameters (P, D, Q, S), and retrain the ARIMA model using the extended dataset to ensure that the model can learn the impact of climate change and surface changes; The enhanced ARIMA model was cross-validated to check its prediction accuracy, especially its performance in reflecting changes in complex environmental conditions. The Akaike Information Criterion (AIC) was used again to evaluate the goodness of fit of the model and select the optimal model parameter configuration. Based on the prediction results and real-time environmental monitoring data, the surveying and mapping data collection strategy and analysis parameters are dynamically adjusted to respond to environmental changes.

[0016] Furthermore, the adjustment module specifically includes: Adjust the frequency of data collection: Adjust the frequency of data collection based on the severity or speed of predicted results and environmental changes.

[0017] Adjustment of collection scope: Adjust the spatial scope of surveying and mapping based on regional environmental changes shown by environmental monitoring data.

[0018] Data resolution: In areas that require detailed monitoring, the resolution of data collection is increased to capture changes more accurately; Analytical algorithm adjustment: Adjust the parameters used in data analysis, including changing the parameters of the statistical model, or adjusting the model's feature selection and weights to cope with changing data characteristics and environmental conditions.

[0019] Beneficial effects of the present invention: The present invention, the cross-domain big data fusion analysis and intelligent prediction system, integrates multi-source mapping data from geographic information systems (GIS), satellite remote sensing, drone measurements and ground stations, and combines them with real-time environmental monitoring data, which significantly improves the environmental adaptability and accuracy of the prediction model. By collecting climate change and surface change data in real time, the prediction model can be dynamically adjusted to more accurately reflect current and future environmental conditions. This comprehensive prediction method can capture environmental changes in a timely manner, thereby providing more accurate support for decisions related to geographic information.

[0020] The present invention realizes dynamic adjustment of the frequency, scope and resolution of surveying and mapping data collection. These adjustments are based on the prediction results provided by the ARIMA model and the real-time environmental monitoring data. This flexible data collection strategy not only ensures the efficiency and economy of the data collection process, but also improves the data quality and relevance. Especially under rapidly changing environmental conditions, by automatically adjusting the collection parameters, the system can more effectively respond to sudden geographical events and seasonal changes, and ensure the timeliness and accuracy of the collected data. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 A schematic diagram of system function modules according to an embodiment of the present invention; Figure 2 Schematic diagram of the ARIMA model of an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0024] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0025] like Figure 1-Figure 2 As shown in the figure, the cross-domain big data fusion analysis and intelligent prediction system includes the following modules: Data acquisition module for collecting spatial and temporal data from multiple mapping domains; Data preprocessing module, used to clean, standardize and format the collected data, perform measurement conversion on data from different sources, ensure that all data units are unified (such as converting inches to centimeters) to eliminate the impact of measurement differences. Standardize the data to a specific range or distribution (such as 0-1 range or standard normal distribution) to facilitate subsequent data processing and analysis. Adjust the data structure to meet the input requirements of subsequent processing modules, such as converting flat data into structured or semi-structured formats; Data integration module, used to integrate and unify data from different surveying and mapping domains, and provide a unified data access interface; Intelligent prediction module, which uses machine learning algorithms to perform trend analysis and prediction on integrated data; The monitoring and perception module collects environmental condition data (climate change, surface change) in real time through an environmental sensor network and combines it with surveying and mapping data to enhance the environmental adaptability and accuracy of the prediction; The adjustment module automatically adjusts the data collection strategy and analysis parameters according to the intelligent prediction and monitoring results, and optimizes the data processing process.

[0026] The multi-surveying and mapping domains include geographic information system (GIS), satellite remote sensing, drone surveying and ground survey stations. The data acquisition modules specifically include: Interface integration mechanism: It has a data interface compatible with multiple surveying and mapping domains and can simultaneously receive multiple data formats, including vector data, raster data, and real-time streaming data; Data filtering and preliminary classification: During the data collection process, implement automated data quality assessment and filtering to eliminate erroneous or incomplete data input, and perform preliminary classification according to data source and type.

[0027] The data integration module includes automatic alignment of spatial data from different sources and at different scales, using geocoding and coordinate conversion techniques to ensure that all data are represented in the same spatial reference, and using time synchronization technology to ensure that data from different surveying and mapping domains have consistent time tags to obtain time series data.

[0028] The machine learning algorithm in the intelligent prediction module adopts the ARIMA model (autoregressive integrated moving average model), which includes: Apply difference processing to the time series data after integration by the data integration module to achieve stationarity requirements, especially for those data that show high autocorrelation and seasonal fluctuations. For seasonally varying data, such as satellite data affected by cloud cover in a particular season, seasonal adjustment methods are applied to eliminate periodic effects; Model parameter identification and optimization: Based on the ACF and PACF plots of the integrated data, identify the ARIMA model parameters p, d, q and seasonal parameters P, D, Q, S, use historical data sets to train the ARIMA model, and select the optimal model using the Akaike Information Criterion AIC; The optimized ARIMA model is used to predict future trends in geographic information changes, including ground deformation and vegetation cover changes. The prediction results are input into the adjustment module to automatically adjust the data collection strategy and analysis parameters to optimize the data processing process.

[0029] By comparing the actual survey data with the predicted data, the prediction accuracy is evaluated, and according to the deviation between the predicted results and the actual data, the parameters of the ARIMA model are adjusted to improve the accuracy and reliability of the prediction.

[0030] Difference processing is used to make non-stationary time series data stationary, as follows: First-order difference: Remove the linear trend of the data and is calculated as: in, is a time series at a point in time The observed value of is the value after the first-order difference. If the data is still not stable after the first-order difference, the second-order difference is performed: ; The seasonal adjustment method is based on seasonal differences and is used to process time series data that show seasonal changes. It removes changes within a fixed seasonal cycle and makes the series stable on the seasonal cycle. Suppose the seasonal cycle is (For annual periodic data, Take 12), the seasonal difference is expressed as: ,in, is the value after seasonal difference.

[0031] In the present invention, the steps of applying the difference and seasonal difference are as follows: Time series evaluation after data integration: First, evaluate the integrated time series data to determine whether the data shows obvious trends and / or seasonal fluctuations.

[0032] Apply differencing: Apply first-order and, if necessary, second-order differencing to the data until the time series exhibits stationary statistical properties (e.g., confirm stationarity via an ADF test).

[0033] Apply seasonal differencing: If the time series shows seasonal fluctuations, apply seasonal differencing, especially when the seasonal effect is very obvious in the original data.

[0034] Model fitting: After necessary difference processing, the time series data should be suitable for fitting and forecasting the ARIMA model.

[0035] The autocorrelation function ACF shows the correlation between the time series and itself at different lags, which is used to identify the order q of the MA model; the partial autocorrelation function PACF graph shows the correlation between the time series and itself under the influence of given other lag values, which is used to identify the order p of the AR model; The identification of non-seasonal parameters p, d, q is as follows: p: Observe the PACF plot and find the lag number with obvious truncation (sudden drop to 0 and close to 0), which indicates the order of the AR part; q: Observe the ACF plot and find the number of lags that are obviously truncated, indicating the order of the MA part; d: The difference order d is determined by the ADF test until the time series becomes stationary; The identification of seasonal parameters P, D, Q, S is as follows: S: seasonal period, determined according to the periodicity of the dataset, such as annually, quarterly, or monthly; P, Q: Same as non-seasonal parameters, determined from seasonal PACF and ACF plots, respectively; D: seasonal difference order, determined by seasonal stationarity test; The ARIMA model uses historical data sets and is trained based on determined model parameters to build an ARIMA model, including dividing the data into training and test sets, applying differencing and seasonal differencing processing, and ensuring the stationarity of the model input data.

[0036] Akaike Information Criterion AIC is expressed as: ,in, the number of model parameters, It is the maximum likelihood estimate of the model. During the model training process, the AIC value is calculated for different parameter combinations, and the model with the lowest AIC value is selected as the optimal model, because a lower AIC value indicates a better balance between the model's fitting quality and complexity.

[0037] The ARIMA model, or Autoregressive Integrated Moving Average Model, is a statistical model used for time series data analysis and forecasting. It combines the characteristics of autoregression (AR), difference (I), and moving average (MA) to adapt to the characteristics of various time series data, especially non-seasonal data. In addition, when the data exhibits seasonal changes, the extended form of seasonal ARIMA (SARIMA) is also often used.

[0038] The ARIMA model is expressed as ARIMA(p, d, q), where p is the order of the autoregressive term, which indicates the number of past values ​​used in the model to predict the current value, d is the number of differences, which indicates the number of non-seasonal differences required to make the time series stationary, and q is the order of the moving average term, which indicates the number of past forecast errors used to predict the current value. The model is expressed as: ,in, It's time The original sequence of is the parameter of the autoregressive term, is the parameter of the moving average term, is the lag operator , is the white noise error term; The seasonal ARIMA is expressed as SARIMA. When the time series data show obvious seasonal fluctuations, SARIMA is used, which is expressed as , where P, D, Q represent the orders of seasonal autoregressive term, seasonal difference and seasonal moving average term respectively, and S represents the seasonal length of the time series. SARIMA is expressed as: ,in, and are the parameters of the seasonal autoregressive and seasonal moving average terms, is the seasonal lag operator, , for or .

[0039] The monitoring and perception module specifically includes: Environmental data integration: real-time environmental condition data collected from environmental sensor networks are synchronized in time and space with surveying and mapping data to ensure that all data items are aligned before analysis. Environmental condition data, including climate factors (temperature, humidity, rainfall) and surface characteristics (vegetation coverage, terrain changes), are feature extracted and converted into a format that can be used for model input; The environmental characteristic data are fused with the integrated surveying and mapping data to create an extended dataset that contains not only time series data but also environmental variables to provide comprehensive prediction factors. Differentiation and transformation processing are applied to the extended dataset to ensure that the data meets the stationarity requirements of the ARIMA model. ARIMA model enhancement: Adjust the parameters (p, d, q) of the ARIMA model based on the extended dataset, consider the impact of environmental factors on the model dynamics, re-evaluate the seasonal parameters (P, D, Q, S), and retrain the ARIMA model using the extended dataset to ensure that the model can learn the impact of climate change and surface changes; The enhanced ARIMA model was cross-validated to check its prediction accuracy, especially its performance in reflecting changes in complex environmental conditions. The Akaike Information Criterion (AIC) was used again to evaluate the goodness of fit of the model and select the optimal model parameter configuration. Based on the prediction results and real-time environmental monitoring data, the surveying and mapping data collection strategy and analysis parameters are dynamically adjusted to respond to environmental changes.

[0040] The adjustment module specifically includes: Integration of prediction and real-time data: Combine the prediction results of the ARIMA model with real-time environmental monitoring data to create a comprehensive data view for decision support. By comparing the prediction results of the model with the latest environmental monitoring data actually collected, the accuracy and reliability of the prediction model can be evaluated. Adjust the frequency of data collection: Adjust the frequency of data collection based on the severity or speed of predicted results and environmental changes. For example, if the forecast model predicts that a certain area will undergo rapid surface changes, the frequency of data collection in that area can be increased to capture details and changes.

[0041] Adjustment of collection scope: Adjust the spatial scope of surveying and mapping based on regional environmental changes shown by environmental monitoring data, for example, expanding coverage of areas severely affected by climate, or focusing on key areas of change.

[0042] Data resolution: In areas that require detailed monitoring, the resolution of data collection is increased to capture changes more accurately; Analytical algorithm adjustment: Adjust the parameters used in data analysis, including changing the parameters of the statistical model, or adjusting the feature selection and weights of the model to cope with the changing data characteristics and environmental conditions; Implement automated systems that can automatically adjust acquisition strategies and analysis parameters based on real-time monitoring data and forecasts, including using controllers to decide when and where to increase mapping frequency or adjust analysis methods.

[0043] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present invention is limited to these examples. Under the concept of the present invention, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity.

[0044] The present invention is intended to cover all such substitutions, modifications and variations that fall within the broad scope of the claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. Cross-domain big data fusion analysis and intelligent prediction system, characterized by: Includes the following modules: Data acquisition module for collecting spatial and temporal data from multiple mapping domains; Data preprocessing module, used to clean, standardize and format the collected data; Data integration module, used to integrate and unify data from different surveying and mapping domains, and provide a unified data access interface; Intelligent prediction module, which uses machine learning algorithms to perform trend analysis and prediction on integrated data; The monitoring and perception module collects environmental condition data in real time through an environmental sensor network and combines it with surveying and mapping data to enhance the environmental adaptability and accuracy of the prediction; The adjustment module automatically adjusts the data collection strategy and analysis parameters according to the intelligent prediction and monitoring results, and optimizes the data processing process.

2. The cross-domain big data fusion analysis and intelligent prediction system according to claim 1 is characterized in that: The multi-surveying and mapping domains include geographic information system GIS, satellite remote sensing, drone measurement and ground survey stations, and the data acquisition module specifically includes: Interface integration mechanism: It has a data interface compatible with multiple surveying and mapping domains and can simultaneously receive multiple data formats, including vector data, raster data, and real-time streaming data; Data filtering and preliminary classification: During the data collection process, implement automated data quality assessment and filtering to eliminate erroneous or incomplete data input, and perform preliminary classification according to data source and type.

3. The cross-domain big data fusion analysis and intelligent prediction system according to claim 1 is characterized in that: The data integration module includes automatically aligning spatial data from different sources and at different scales, using geocoding and coordinate conversion techniques to ensure that all data are represented in the same spatial reference, and using time synchronization technology to ensure that data from different surveying and mapping domains have consistent time tags to obtain time series data.

4. The cross-domain big data fusion analysis and intelligent prediction system according to claim 3 is characterized in that: The machine learning algorithm in the intelligent prediction module adopts the ARIMA model, which specifically includes: Apply differential processing to the time series data after integration by the data integration module to achieve stationarity requirements, and apply seasonal adjustment methods to seasonally changing data to eliminate cyclical effects; Model parameter identification and optimization: Based on the ACF and PACF plots of the integrated data, identify the ARIMA model parameters p, d, q and seasonal parameters P, D, Q, S, use historical data sets to train the ARIMA model, and select the optimal model using the Akaike Information Criterion AIC; The optimized ARIMA model is used to predict future trends in geographic information changes, including ground deformation and vegetation cover changes. The prediction results are input into the adjustment module to automatically adjust the data collection strategy and analysis parameters to optimize the data processing process.

5. The cross-domain big data fusion analysis and intelligent prediction system according to claim 4 is characterized in that: The difference processing is used to make the non-stationary time series data stationary, as follows: First-order difference: Remove the linear trend of the data and is calculated as: in, is a time series at a point in time The observed value of is the value after the first-order difference. If the data is still not stable after the first-order difference, the second-order difference is performed: ； The seasonal adjustment method is based on seasonal differences and is used to process time series data that show seasonal changes, removing changes within a fixed seasonal cycle and making the series stable on the seasonal cycle. Suppose the seasonal cycle is , the seasonal difference is expressed as: ,in, is the value after seasonal difference.

6. The cross-domain big data fusion analysis and intelligent prediction system according to claim 4 is characterized in that: The autocorrelation function ACF shows the correlation between the time series and itself at different lags, which is used to identify the order q of the MA model; the partial autocorrelation function PACF graph shows the correlation between the time series and itself under the influence of given other lag values, which is used to identify the order p of the AR model; The identification of the non-seasonal parameters p, d, q is as follows: p: Observe the PACF plot and find the number of lags that are obviously truncated, which indicates the order of the AR part; q: Observe the ACF plot and find the number of lags that are obviously truncated, indicating the order of the MA part; d: The difference order d is determined by the ADF test until the time series becomes stationary; The identification of seasonal parameters P, D, Q, S is as follows: S: seasonal period, determined according to the periodicity of the data set; P, Q: Same as non-seasonal parameters, determined from seasonal PACF and ACF plots, respectively; D: seasonal difference order, determined by seasonal stationarity test; The ARIMA model uses a historical data set and is trained according to determined model parameters to construct the ARIMA model, including dividing the data into a training set and a test set, applying difference and seasonal difference processing, and ensuring the stationarity of the model input data.

7. The cross-domain big data fusion analysis and intelligent prediction system according to claim 4 is characterized in that: The Akaike Information Criterion AIC is expressed as: ,in, the number of model parameters, It is the maximum likelihood estimate of the model. During the model training process, the AIC values ​​are calculated for different parameter combinations, and the model with the lowest AIC value is selected as the optimal model.

8. The cross-domain big data fusion analysis and intelligent prediction system according to claim 4 is characterized in that: The ARIMA model is expressed as ARIMA(p, d, q), where p is the order of the autoregressive term, which indicates the number of past values ​​used in the model to predict the current value, d is the number of differences, which indicates the number of non-seasonal differences required to make the time series stationary, and q is the order of the moving average term, which indicates the number of past forecast errors used to predict the current value. The model is expressed as: ,in, It's time The original sequence of is the parameter of the autoregressive term, is the parameter of the moving average term, is the lag operator , is the white noise error term; The seasonal ARIMA is expressed as SARIMA. SARIMA is used when time series data show obvious seasonal fluctuations. , where P, D, Q represent the orders of seasonal autoregressive term, seasonal difference and seasonal moving average term respectively, and S represents the seasonal length of the time series. SARIMA is expressed as: ,in, and are the parameters of the seasonal autoregressive and seasonal moving average terms, is the seasonal lag operator, , for or .

9. The cross-domain big data fusion analysis and intelligent prediction system according to claim 4 is characterized in that: The monitoring and sensing module specifically includes: Environmental data integration: real-time environmental condition data collected from environmental sensor networks are synchronized in time and space with surveying and mapping data to ensure that all data items are aligned before analysis, and feature extraction is performed on environmental condition data, including climate factors and surface features, and converted into a format that can be used for model input; The environmental characteristic data are fused with the integrated surveying and mapping data to create an extended dataset that contains not only time series data but also environmental variables to provide comprehensive prediction factors. Differentiation and transformation processing are applied to the extended dataset to ensure that the data meets the stationarity requirements of the ARIMA model. ARIMA model enhancement: Adjust the parameters p, d, q of the ARIMA model based on the extended data set, consider the impact of environmental factors on the model dynamics, re-evaluate the seasonal parameters P, D, Q, S, and retrain the ARIMA model using the extended data set to ensure that the model can learn the impact of climate change and surface changes; The enhanced ARIMA model was cross-validated to check its prediction accuracy, and the Akaike Information Criterion (AIC) was used again to evaluate the goodness of fit of the model and select the optimal model parameter configuration; Based on the prediction results and real-time environmental monitoring data, the surveying and mapping data collection strategy and analysis parameters are dynamically adjusted to respond to environmental changes.

10. The cross-domain big data fusion analysis and intelligent prediction system according to claim 9 is characterized in that: The adjustment module specifically includes: Adjust the frequency of data collection: Adjust the frequency of data collection based on the severity or speed of the forecast results and environmental changes; Adjustment of collection scope: Adjust the spatial scope of surveying and mapping according to regional environmental changes shown by environmental monitoring data; Data resolution: In areas that require detailed monitoring, the resolution of data collection is increased to capture changes more accurately; Analytical algorithm adjustment: Adjust the parameters used in data analysis, including changing the parameters of the statistical model, or adjusting the model's feature selection and weights to cope with changing data characteristics and environmental conditions.

Images