Ecological environmental protection monitoring method based on user behaviors

Abstract

The invention relates to an ecological environmental protection monitoring method based on user behaviors. The method comprises the following steps: collecting multi-source data from a home office travel scene; wherein the multi-source data comprises an electric power water resource fuel consumption record; detecting a high-consumption event for the multi-source data, and determining an event association sequence; acquiring key behavior characteristics based on the event association sequence; generating a scene specific association model according to the key behavior feature set, and determining an optimization rule; and performing iterative matching on the real-time acquired data by adopting an optimization rule to obtain personalized environmental protection suggestions. Through deep fusion of data synchronization and behavior analysis, an efficient and personalized resource saving scheme is provided, and the environmental protection efficiency and the intelligent level under multiple scenes are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of ecological and environmental monitoring technology, and in particular to an ecological and environmental monitoring method based on user behavior. Background Technology

[0002] Ecological and environmental monitoring, as a crucial research area for addressing excessive resource consumption and environmental degradation, aims to reveal the impact of human activities on the ecological environment through scientific methods, providing key support for sustainable development. Accurately monitoring the relationship between user behavior and resource consumption has become central to promoting green lifestyles and policy optimization. However, existing methods have significant technical limitations, making it difficult to meet the dynamic and personalized needs of environmental monitoring. Many current monitoring methods rely primarily on static data collection or single-scenario analysis, ignoring the complexity and dynamic changes of user behavior across different scenarios. For example, household electricity consumption monitoring often only records the total amount, failing to deeply analyze how specific user behaviors affect consumption patterns. This method cannot capture the differences in user behavior patterns in scenarios such as home, office, and travel, resulting in a lack of targeted analysis results and difficulty in supporting personalized environmental optimization recommendations. A deeper technical challenge lies in how to achieve real-time correlation analysis between user behavior and resource consumption through IoT devices. IoT devices need to collect high-frequency, heterogeneous data in multiple scenarios, such as electricity, water resources, and fuel consumption. These data come from diverse sources and are in various formats, making integration difficult. The real-time nature of data acquisition requires devices with high-efficiency processing capabilities. However, in practice, communication delays between devices and inconsistencies in data processing often lead to information distortion. For example, a household may use multiple smart devices simultaneously to record electricity consumption, but due to poor data synchronization between devices, it is difficult to accurately determine the contribution of a specific behavior (such as prolonged use of air conditioning) to total electricity consumption. This data integration and real-time issue further exacerbates another technical challenge: the accurate identification of behavioral characteristics. User behavior is highly personalized; for example, in the same scenario, different users' habits of using appliances may lead to drastically different resource consumption patterns. Existing technologies, when analyzing behavioral characteristics, often cannot effectively distinguish which behaviors directly lead to resource waste and which are normal uses, thus making it difficult to provide targeted environmental protection recommendations. Therefore, how to achieve real-time and accurate correlation analysis between user behavior and resource consumption through IoT devices in multiple scenarios, and accurately identify key behavioral characteristics based on this, has become a key issue driving breakthroughs in ecological and environmental monitoring technology. Summary of the Invention

[0003] The purpose of this invention is to propose an ecological and environmental monitoring method based on user behavior to solve the problems existing in the prior art.

[0004] To achieve the above objectives, the present invention provides the following solution: User behavior-based ecological and environmental monitoring methods include: Collect multi-source data from home office and travel scenarios; wherein, the multi-source data includes: electricity, water resources, and fuel consumption records; For the multi-source data, high-consumption events are detected, and the event correlation sequence is determined; Based on the event association sequence, key behavioral features are obtained; Generate a scenario-specific association model based on the key behavioral feature set and determine the optimization rules; By employing optimized rules to iteratively match real-time collected data, personalized environmental protection recommendations are obtained.

[0005] Optionally, data collected from multiple sources in home office / travel scenarios includes: High-frequency heterogeneous data are collected from home, office and travel scenarios through IoT devices to obtain multi-source raw datasets; The network formats the original datasets from multiple sources to generate a dataset with a unified structure. If there are missing data on electricity consumption, water consumption, and fuel consumption in the unified structure dataset, the missing values ​​are filled in by linear interpolation to obtain the complete dataset. A time synchronization mechanism is used to calibrate the communication delay between devices in the complete dataset, generating a synchronized dataset.

[0006] Optionally, for the multi-source data, detecting high-consumption events and determining the event association sequence includes: If the power consumption records in the multi-source data exceed a preset threshold, then high-consumption events are marked by comparing time series data to generate a set of high-consumption events; Based on the set of high-consumption events, extract the event occurrence time and device identification information to generate an event time identification dataset; By using an event time-identified dataset, a time series analysis algorithm is employed to calculate the correlation between the event occurrence time and the data collection frequency, thereby determining the event association sequence.

[0007] Optionally, based on the event association sequence, obtaining key behavioral features includes: Sequence analysis was used to cluster user behavior patterns in event-related sequences to obtain behavior cluster groups. The correlation coefficient is calculated by matching behavioral clusters with resource consumption records to determine the strength of the association between behavioral consumption. If the correlation strength of the behavior consumption is higher than the preset threshold, the corresponding behavior features are extracted to obtain key behavior features.

[0008] Optionally, using sequence analysis to cluster user behavior patterns in event-related sequences includes: Sequence analysis is used to extract the temporal order and behavioral frequency from the event association sequence to obtain a behavioral feature set; If the temporal order of the behavioral feature set meets the preset time interval threshold, then the feature extraction method is used to generate a behavioral pattern feature vector. The behavioral pattern feature vectors are grouped using the K-means clustering algorithm to obtain preliminary clustering groups. If the event association strength of the initial clustering group is higher than the preset threshold, then pattern recognition technology is used to optimize the grouping to obtain an optimized behavior pattern group. Based on the optimized behavior pattern group, sequence analysis was used to verify the consistency of behavior frequency within the group, and the final behavior pattern group was obtained. By using the final behavior pattern group, a clustering algorithm is used to update the group boundaries, resulting in a stable behavior cluster group.

[0009] Optionally, the correlation coefficient is calculated by matching behavioral clusters with resource consumption records to determine the strength of the association between behavioral consumption and records. Data is aligned by behavioral clustering groups and resource consumption records, and a timestamp matching method is used to obtain the aligned dataset; If the time interval of the aligned dataset meets the preset threshold, then a behavioral feature vector is generated through behavioral pattern feature extraction. Based on the distribution of behavioral feature vectors and consumption data, the Pearson correlation coefficient method is used to calculate the correlation coefficient between behavior and consumption, and to determine the strength of the association between behavior and consumption.

[0010] Optionally, a scenario-specific association model is generated based on the key behavioral feature set, and the optimization rules are determined, including: Based on the key behavioral feature set, cluster analysis is used to determine scene associations; The correlation strength is calculated based on the scene association, and the correlation strength value is obtained by using the Pearson correlation coefficient method. If the association strength value is greater than the preset threshold, an association model is constructed based on the association strength to obtain a scenario-specific association model; The optimization objective is extracted from the scene association model, and the optimization rules are determined using the decision tree algorithm.

[0011] Optionally, optimization rules are used to iteratively match the real-time collected data to obtain personalized environmental protection suggestions, including: Based on the structured environmental data collected in real time, preliminary matching is performed using optimized rules to determine the initial matching results; If the rule matching accuracy of the initial matching result is lower than the preset threshold, the rule set is optimized using the random forest algorithm to obtain the updated optimized rules. The updated optimization rules are used to perform iterative matching on structured environment data to generate personalized suggestions.

[0012] The beneficial effects of this invention are as follows: This invention proposes an eco-environmental monitoring method based on user behavior. It collects high-frequency heterogeneous data from home, office, and travel scenarios using IoT devices, integrates electricity, water, and fuel consumption records to construct a multi-source dataset, and calibrates communication latency between devices through a time synchronization mechanism to form a synchronized dataset. This addresses the business problem of monitoring resource consumption and analyzing behavioral correlations across multiple scenarios. The invention labels high-consumption events, extracts event correlation sequences, clusters user behavior patterns through sequence analysis, calculates the correlation coefficient between behavior and resource consumption, identifies high-strength correlated behavioral characteristics, and generates scenario-specific correlation models and optimization rule sets. The invention uses the optimization rule set to iteratively match real-time data, generating personalized environmental protection suggestion sequences, achieving accurate monitoring and optimized management of resource consumption across scenarios. Its core technological effect lies in providing efficient and personalized resource-saving solutions through the deep integration of data synchronization and behavioral analysis, significantly improving environmental efficiency and intelligence levels across multiple scenarios. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a schematic diagram of the process of the ecological and environmental monitoring method based on user behavior according to an embodiment of the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] like Figure 1 As shown, this embodiment proposes an ecological and environmental monitoring method based on user behavior, including: Collect multi-source data from home office and travel scenarios; wherein, the multi-source data includes: electricity, water resources, and fuel consumption records; For the multi-source data, high-consumption events are detected, and the event correlation sequence is determined; Based on the event association sequence, key behavioral features are obtained; Generate a scenario-specific association model based on the key behavioral feature set and determine the optimization rules; By employing optimized rules to iteratively match real-time collected data, personalized environmental protection recommendations are obtained.

[0018] Furthermore, the collection of multi-source data in home office and travel scenarios includes: High-frequency heterogeneous data are collected from home, office and travel scenarios through IoT devices to obtain multi-source raw datasets; The network formats the original datasets from multiple sources to generate a dataset with a unified structure. If there are missing data on electricity consumption, water consumption, and fuel consumption in the unified structure dataset, the missing values ​​are filled in by linear interpolation to obtain the complete dataset. A time synchronization mechanism is used to calibrate the communication delay between devices in the complete dataset, generating a synchronized dataset.

[0019] Specifically, in this embodiment, in the home office and travel scenario, IoT devices deployed in smart homes and vehicles, such as smart meters, water meters, gas sensors, and GPS modules, are used to collect high-frequency heterogeneous data at a sampling frequency of one minute. For example, the power consumption record shows that the user's average power consumption during office work is 150W, with a peak of 300W, water consumption is 2.5L per day for drinking and sanitation, and fuel consumption is 0.08L of diesel per kilometer for commuting vehicles. The data is uploaded to the cloud server in real time via wireless transmission protocols such as MQTT, forming a multi-source raw dataset.

[0020] First, data cleaning algorithms are used to remove outliers. For example, the Z-score method is used to calculate the Z score of each electricity record. If |Z|>3, it is considered noise and removed, with a removal rate of 5%. Then, the heterogeneous features are processed by principal component analysis (PCA) algorithm to reduce the dimensionality and fuse the electricity, water and fuel data into 3-dimensional principal components, retaining 95% of the variance explained, so as to reduce the computational complexity and improve the efficiency of subsequent analysis.

[0021] A time synchronization mechanism is used to calibrate the communication latency between devices in the multi-source raw dataset, generating a synchronized dataset. Timestamp information from each device is extracted from the synchronized dataset to generate a timestamp-aligned dataset. If data points in the timestamp-aligned dataset have time deviations exceeding a preset threshold, linear interpolation is used to correct the deviations, resulting in a time-consistent dataset.

[0022] Furthermore, for the multi-source data, high-consumption events are detected, and the event association sequence is determined to include: If the power consumption records in the multi-source data exceed a preset threshold, then high-consumption events are marked by comparing time series data to generate a set of high-consumption events; Based on the set of high-consumption events, extract the event occurrence time and device identification information to generate an event time identification dataset; By using an event time-identified dataset, a time series analysis algorithm is employed to calculate the correlation between the event occurrence time and the data collection frequency, thereby determining the event association sequence.

[0023] Specifically, in this embodiment, for example, during the multi-source data acquisition phase, current and voltage data are collected in real time from three industrial devices D, E, and F via an IoT sensor network to form a raw dataset. Each device records 15 seconds of data at a sampling rate of 2kHz, generating a total of 45,000 sampling points. Among them, device D has a peak current of 120.5A and an average voltage of 220.3V, device E has a peak current of 118.7A and an average voltage of 221.1V, and device F has a peak current of 119.2A and an average voltage of 219.8V. Due to clock skew between devices, the timestamps are offset, with preliminary analysis showing an average offset of 200ms. The cross-correlation function between datasets, R_{DE}(τ)=∑[D(t)·E(t+τ)] / N, is calculated, where N is the number of sampling points and τ is the delay. The peak position τ_max=0.15s indicates that D leads E by 150ms. Similarly, the offset values ​​are obtained from R_{DF}(τ_max=0.22s) and R_{EF}(τ_max=0.07s), forming a time deviation matrix ΔT=[0,0.15,0.22;-0.15,0,-0.07;-0.22,0.07,0]. To calibrate the communication delay, the PTP protocol is used, sending 100 synchronization messages per second. The average round-trip time (RTT) is measured to be 20ms, with a standard deviation σ=2ms. Using the extended Kalman filter algorithm, the state equation is x_k = f(x_{k-1}) + w_k, the observation equation is z_k = h(x_k) + v_k, the initial state is x_0 = 18ms, the covariance is P_0 = 8, the process noise is Q = 0.5, and the observation noise is R = 3. After 150 iterations, the calibration delays are d_D = 19.2ms, d_E = 20.8ms, and d_F = 20.1ms. Combined with the bias matrix, the calibration factor β_i = 1 / (1 + d_i / 1000) is calculated, and the timestamp t'_i = t_i - β_i·t_i is adjusted. Finally, the time axis is aligned using a quadratic interpolation algorithm, with the formula f(t) = a_0 + a_1t + a_2t^2, where the coefficients are fitted using the least squares method, and the clock is resampled to a uniform 2kHz clock. If the power consumption P=UI exceeds the threshold of 10000W, it is marked as a high consumption event. The correlation of the event sequence is analyzed, and the cross-correlation coefficient ρ between P and current I is calculated to be 0.95. It is determined that the high consumption event is associated with the current mutation. The sequence length is 500ms.

[0024] Furthermore, based on the event association sequence, key behavioral features are obtained, including: Sequence analysis was used to cluster user behavior patterns in event-related sequences to obtain behavior cluster groups. The correlation coefficient is calculated by matching behavioral clusters with resource consumption records to determine the strength of the association between behavioral consumption. If the correlation strength of the behavior consumption is higher than the preset threshold, the corresponding behavior features are extracted to obtain key behavior features.

[0025] Furthermore, the use of sequence analysis to cluster user behavior patterns in event-related sequences includes: Sequence analysis is used to extract the temporal order and behavioral frequency from the event association sequence to obtain a behavioral feature set; If the temporal order of the behavioral feature set meets the preset time interval threshold, then the feature extraction method is used to generate a behavioral pattern feature vector. The behavioral pattern feature vectors are grouped using the K-means clustering algorithm to obtain preliminary clustering groups. If the event association strength of the initial clustering group is higher than the preset threshold, then pattern recognition technology is used to optimize the grouping to obtain an optimized behavior pattern group. Based on the optimized behavior pattern group, sequence analysis was used to verify the consistency of behavior frequency within the group, and the final behavior pattern group was obtained. By using the final behavior pattern group, a clustering algorithm is used to update the group boundaries, resulting in a stable behavior cluster group.

[0026] Furthermore, by matching behavioral clusters with resource consumption records to calculate correlation coefficients, the strength of the association between behavioral consumption is determined, including: Data is aligned by behavioral clustering groups and resource consumption records, and a timestamp matching method is used to obtain the aligned dataset; If the time interval of the aligned dataset meets the preset threshold, then a behavioral feature vector is generated through behavioral pattern feature extraction. Based on the distribution of behavioral feature vectors and consumption data, the Pearson correlation coefficient method is used to calculate the correlation coefficient between behavior and consumption, and to determine the strength of the association between behavior and consumption.

[0027] Furthermore, a scenario-specific association model is generated based on the key behavioral feature set, and the optimization rules are determined, including: Based on the key behavioral feature set, cluster analysis is used to determine scene associations; The correlation strength is calculated based on the scene association, and the correlation strength value is obtained by using the Pearson correlation coefficient method. If the association strength value is greater than the preset threshold, an association model is constructed based on the association strength to obtain a scenario-specific association model; The optimization objective is extracted from the scene association model, and the optimization rules are determined using the decision tree algorithm.

[0028] Furthermore, by employing optimized rules to iteratively match the real-time collected data, personalized environmental protection suggestions are obtained, including: Based on the structured environmental data collected in real time, preliminary matching is performed using optimized rules to determine the initial matching results; If the rule matching accuracy of the initial matching result is lower than the preset threshold, the rule set is optimized using the random forest algorithm to obtain the updated optimized rules. The updated optimization rules are used to perform iterative matching on structured environment data to generate personalized suggestions.

[0029] By using a user behavior analysis model and combining it with environmental data input, the personalized suggestion set is sorted and filtered to obtain an environmental protection suggestion sequence. The environmental protection recommendation sequence is dynamically adjusted based on the sequence optimization method to generate the final recommendation sequence; If the final suggested sequence deviates from the dynamic update of the real-time collected data, the dynamic data update is classified using the K-means clustering algorithm to obtain the updated suggested sequence.

[0030] Specifically, in this embodiment, for example, after collecting user vehicle driving data in real time, the system first uses an optimized rule set to preprocess the collected average hourly carbon dioxide emissions of 15.3 kg and particulate matter concentration of 0.045 mg / m³. The data is then divided into three clusters—low, medium, and high—using the K-means clustering algorithm, with cluster centers at emission thresholds of 10 kg, 15 kg, and 20 kg, respectively, to identify abnormal peaks such as sudden emission points exceeding 18 kg, thereby filtering out noisy data to form a clean dataset. Next, an iterative matching process is initiated based on this dataset. The cosine similarity formula cosθ=(A·B) / (||A||·||B||) is used to compare the user's historical environmental behavior vector with the carbon reduction rule vector in the rule set. For example, the user vector [0.8,0.6,0.4] represents daily commuting, diet, and energy usage habits, and its similarity with the rule vector [0.7,0.5,0.9] is 0.92, which is higher than the threshold of 0.85, triggering the first matching to generate preliminary suggestions such as "Switching to electric vehicles for commuting can reduce emissions by 5 kg". In the second iteration, the "Analyzing user request - real-time feedback" is incorporated to adjust the weights. The gradient descent algorithm is used to minimize the loss function L=∑(predicted emission reduction - actual emission reduction)^2 / 2, and the rule weights are updated with a learning rate of 0.01. The weight of the diet rule is increased from 0.3 to 0.45 because the analysis shows that the user's dietary carbon footprint analysis process reveals that weekly meat intake leads to an additional 2.1 kg of emissions. The third iteration further refined the process, applying the Apriori association rule mining algorithm to mine frequent itemsets with a support of min_sup=0.6 and a confidence of min_conf=0.8. It found that the combination of "low-emission commuting + vegetarian days" had a support of 0.75, generating sequential suggestions such as the first step "vegetarian diet for three days this week, reducing emissions by 1.8 kg", the second step "cycling instead of driving on weekends, reducing emissions by 3.2 kg", and the third step "installing a smart thermostat to optimize home energy consumption, reducing emissions by 2.5 kg". The entire sequence is expected to reduce emissions by a total of 7.5 kg, ensuring that the personalized environmental protection path logically progresses from immediate behavior adjustment to long-term habit formation. The effectiveness of the sequence was verified through Monte Carlo simulation. The average emission reduction confidence interval after 1000 iterations was 6.8 to 8.2 kg, forming a closed-loop optimization thinking chain.

[0031] This embodiment collects high-frequency heterogeneous data from home, office, and travel scenarios using IoT devices, integrates electricity, water, and fuel consumption records to construct a multi-source dataset, and calibrates communication latency between devices through a time synchronization mechanism to form a synchronized dataset, thereby solving the business problem of resource consumption monitoring and behavioral correlation analysis in multiple scenarios. This invention marks high-consumption events, extracts event correlation sequences, clusters user behavior patterns through sequence analysis, calculates the correlation coefficient between behavior and resource consumption, identifies high-strength correlated behavioral characteristics, and generates scenario-specific correlation models and optimization rule sets. This invention uses the optimization rule set to iteratively match real-time data to generate personalized environmental protection suggestion sequences, achieving accurate monitoring and optimized management of resource consumption across scenarios. Its core technical effect lies in providing efficient and personalized resource-saving solutions through the deep integration of data synchronization and behavioral analysis, significantly improving environmental efficiency and intelligence levels in multiple scenarios.

[0032] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. An ecological and environmental monitoring method based on user behavior, characterized in that, include: Collect multi-source data from home office and travel scenarios; wherein, the multi-source data includes: electricity, water resources, and fuel consumption records; For the multi-source data, high-consumption events are detected, and the event correlation sequence is determined; Based on the event association sequence, key behavioral features are obtained; Generate a scenario-specific association model based on the key behavioral feature set and determine the optimization rules; By employing optimized rules to iteratively match real-time collected data, personalized environmental protection recommendations are obtained.

2. The ecological and environmental monitoring method based on user behavior according to claim 1, characterized in that, Data collected from multiple sources in home office and travel scenarios includes: High-frequency heterogeneous data are collected from home, office and travel scenarios through IoT devices to obtain multi-source raw datasets; The network formats the original datasets from multiple sources to generate a dataset with a unified structure. If there are missing data on electricity consumption, water consumption, and fuel consumption in the unified structure dataset, the missing values ​​are filled in by linear interpolation to obtain the complete dataset. A time synchronization mechanism is used to calibrate the communication delay between devices in the complete dataset, generating a synchronized dataset.

3. The ecological and environmental monitoring method based on user behavior according to claim 1, characterized in that, For the multi-source data, high-consumption events are detected, and the event association sequences are determined, including: If the power consumption records in the multi-source data exceed a preset threshold, then high-consumption events are marked by comparing time series data to generate a set of high-consumption events; Based on the set of high-consumption events, extract the event occurrence time and device identification information to generate an event time identification dataset; By using an event time-identified dataset, a time series analysis algorithm is employed to calculate the correlation between the event occurrence time and the data collection frequency, thereby determining the event association sequence.

4. The ecological and environmental monitoring method based on user behavior according to claim 1, characterized in that, Based on the event association sequence, key behavioral features are obtained, including: Sequence analysis was used to cluster user behavior patterns in event-related sequences to obtain behavior cluster groups. The correlation coefficient is calculated by matching behavioral clusters with resource consumption records to determine the strength of the association between behavioral consumption. If the correlation strength of the behavior consumption is higher than the preset threshold, the corresponding behavior features are extracted to obtain key behavior features.

5. The ecological and environmental monitoring method based on user behavior according to claim 4, characterized in that, Clustering user behavior patterns in event-related sequences using sequence analysis includes: Sequence analysis is used to extract the temporal order and behavioral frequency from the event association sequence to obtain a behavioral feature set; If the temporal order of the behavioral feature set meets the preset time interval threshold, then the feature extraction method is used to generate a behavioral pattern feature vector. The behavioral pattern feature vectors are grouped using the K-means clustering algorithm to obtain preliminary clustering groups. If the event association strength of the initial clustering group is higher than the preset threshold, then pattern recognition technology is used to optimize the grouping to obtain optimized behavior pattern groups; Based on the optimized behavior pattern group, sequence analysis was used to verify the consistency of behavior frequency within the group, and the final behavior pattern group was obtained. By using the final behavior pattern group, a clustering algorithm is used to update the group boundaries, resulting in a stable behavior cluster group.

6. The ecological and environmental monitoring method based on user behavior according to claim 4, characterized in that, The correlation coefficient is calculated by matching behavioral clusters with resource consumption records to determine the strength of the association between behavioral consumption and other factors. Data is aligned by behavioral clustering groups and resource consumption records, and a timestamp matching method is used to obtain the aligned dataset; If the time interval of the aligned dataset meets the preset threshold, then a behavioral feature vector is generated through behavioral pattern feature extraction. Based on the distribution of behavioral feature vectors and consumption data, the Pearson correlation coefficient method is used to calculate the correlation coefficient between behavior and consumption, and to determine the strength of the association between behavior and consumption.

7. The ecological and environmental monitoring method based on user behavior according to claim 1, characterized in that, Based on the key behavioral feature set, a scenario-specific association model is generated, and the optimization rules are determined, including: Based on the key behavioral feature set, cluster analysis is used to determine scene associations; The correlation strength is calculated based on the scene association, and the correlation strength value is obtained by using the Pearson correlation coefficient method. If the association strength value is greater than the preset threshold, an association model is constructed based on the association strength to obtain a scenario-specific association model; The optimization objective is extracted from the scene association model, and the optimization rules are determined using the decision tree algorithm.

8. The ecological and environmental monitoring method based on user behavior according to claim 1, characterized in that, The real-time collected data is iteratively matched using optimization rules to obtain personalized environmental protection suggestions, including: Based on the structured environmental data collected in real time, preliminary matching is performed using optimized rules to determine the initial matching results; If the rule matching accuracy of the initial matching result is lower than the preset threshold, the rule set is optimized using the random forest algorithm to obtain the updated optimized rules. The updated optimization rules are used to perform iterative matching on structured environment data to generate personalized suggestions.

Images