A Method and System for Optimized Waste Disposal Management Based on Classification Data Mining

Abstract

This invention provides a method and system for optimized waste disposal management based on classification data mining, relating to the field of waste management technology. The method includes: acquiring waste disposal behavior data uploaded by waste disposal collection equipment and generating an original behavior sequence; dividing the data into time periods to generate short-term and long-term classification sequences; performing category difference processing on adjacent time periods and category aggregation calculation on each long-term stage to generate short-term fluctuation indicators and long-term stability indicators; performing cross-sequence association analysis to generate association analysis results and constructing a hierarchical triggering factor including short-term and long-term triggering factors; determining the offset difference, and generating a behavior mutation event when the offset difference between the short-term and long-term triggering factors exceeds a preset association offset threshold; and performing strategy migration processing to generate a migrated waste disposal management strategy. This invention improves the autonomy and accuracy of optimized waste disposal management.

Description

Technical Field

[0001] This invention relates to the field of waste management technology, and in particular to a waste disposal optimization management method and system based on classification data mining. Background Technology

[0002] Existing technologies typically utilize fixed waste disposal management systems. These systems collect real-time data on residents' waste disposal behavior by installing weighing sensors, RFID readers at the disposal openings, or cameras on each bin. Structured data, including disposal volume, time period, and type, is then uploaded to a backend database. The backend system generally uses preset threshold rules or simple clustering algorithms (such as time-based K-means clustering) to analyze this data, generating management results such as waste overflow prediction, waste sorting accuracy assessment, and collection route planning. In actual community deployments, these systems often rely on historical averages or fixed rules to determine disposal behavior patterns. For example, they might use the empirical finding that "the disposal volume of a certain type of waste tends to be higher in the evening" as the basis for subsequent scheduling, thereby achieving basic automation of waste sorting management and collection scheduling.

[0003] However, when faced with scenarios exhibiting significant heterogeneity in waste disposal behavior, the aforementioned existing technologies are prone to problems such as rigid data mining models that struggle to adapt to real-world changes. For example, in some suburban communities, frequent tenant turnover and the waste disposal habits of new residents (such as disposing of large cardboard items in the early morning for three consecutive days) can cause the actual overflow time of trash cans to suddenly arrive earlier. However, rule-based models based on historical averages still rely on the original threshold of "not overflowing before 8 a.m.," failing to dynamically identify this sudden change in behavior, resulting in a large amount of overflowing trash cans as early as 4 a.m. This type of technical deficiency is not simply a matter of "low efficiency or high cost," but rather stems from the difficulty of existing models automatically identifying "abnormal disposal patterns" and "periodic changes." They cannot update behavioral patterns in specific scenarios involving changes in the community's customer demographics, leading to lags and misjudgments in the management system's scheduling output, resulting in practical problems such as waste accumulation, odor spread, and increased frequency of subsequent manual intervention. Summary of the Invention

[0004] The purpose of this invention is to provide a waste disposal optimization management method and system based on classification data mining, which aims to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0006] Firstly, a waste disposal optimization management method based on classification data mining, the method comprising:

[0007] Acquire waste disposal behavior data uploaded by waste disposal collection devices, and generate an original behavior sequence to represent continuous disposal behavior based on the data;

[0008] The original behavior sequence is divided into time periods to generate short-cycle behavior data and long-cycle behavior data, and short-cycle classification sequences and long-cycle classification sequences are constructed respectively.

[0009] Based on the short-cycle classification sequence and the long-cycle classification sequence, the category difference processing of adjacent time periods and the category aggregation calculation of each long-cycle stage are performed to generate short-cycle fluctuation index and long-cycle stability index.

[0010] Cross-sequence correlation analysis is performed based on short-cycle volatility indicators and long-cycle stability indicators to generate correlation analysis results, and a hierarchical triggering factor including short-cycle triggering factors and long-cycle triggering factors is constructed based on these results.

[0011] The offset difference is determined based on the hierarchical triggering factor. When the offset difference between the short-cycle triggering factor and the long-cycle triggering factor exceeds the preset associated offset threshold, a behavioral mutation event is generated.

[0012] Based on the behavioral mutation event, the strategy migration process is executed to generate the migrated waste disposal management strategy, which includes an updated waste bin overflow prediction window, an adjusted waste collection priority, and a reset waste disposal classification correction prompt.

[0013] Execute waste disposal optimization management operations according to the migrated waste disposal management strategy, and output the execution results to the waste disposal management terminal.

[0014] Preferably, based on the short-period classification sequence and the long-period classification sequence, category difference processing is performed on adjacent time periods, and category aggregation calculation is performed on each long-period stage to generate short-period volatility indicators and long-period stability indicators, including:

[0015] The short-period classification sequence is segmented according to a fixed-length short-period analysis window. Within each analysis window, the directional comparison, continuity judgment and amplitude accumulation of the category changes in adjacent time periods are performed to generate short-period difference results within the short-period window.

[0016] Consistency detection is performed on the category change patterns of multiple consecutive analysis windows based on the short-cycle difference results, and the consistency detection results are transformed into short-cycle fluctuation indicators.

[0017] Long-period classification sequences are grouped according to long-period stages. Within each long-period group, long-period clustering results are generated based on the frequency of category occurrence, duration, and stability of the dominant category within the stage. Long-period stability indices are then calculated based on the long-period clustering results.

[0018] Preferably, cross-series correlation analysis is performed based on short-cycle volatility indicators and long-cycle stability indicators to generate correlation analysis results, and a hierarchical triggering factor system is constructed based on these results, including short-cycle triggering factors and long-cycle triggering factors, comprising:

[0019] The short-cycle volatility indicator and the long-cycle stability indicator are time-aligned, and the trend difference value between the two is calculated within the common time period after alignment. The initial correlation result is generated based on the trend difference value.

[0020] Based on the trend difference values ​​in the primary correlation results, the directionality, cumulativeity, and persistence of the difference in the alignment time period are segmented for analysis, generating basic values ​​of short-term triggering factors to describe short-term burst behavior and basic values ​​of long-term triggering factors to describe long-term offset behavior.

[0021] Based on the structural differences between the base values ​​of short-cycle trigger factors and long-cycle trigger factors, the two are hierarchically reorganized, with the short-cycle trigger factor placed in the rapid response layer and the long-cycle trigger factor placed in the trend identification layer, forming a layered trigger factor with two layers.

[0022] Preferably, the offset difference is determined based on the hierarchical triggering factor. When the offset difference between the short-cycle triggering factor and the long-cycle triggering factor exceeds a preset correlation offset threshold, a behavioral mutation event is generated, including:

[0023] Offset source filtering is performed on the short-cycle trigger factors and long-cycle trigger factors of the layered trigger factors respectively. The instantaneous changes caused by a single anomaly in the short-cycle trigger factors are removed, and the unstable offsets caused by short-term fluctuations in the long-cycle trigger factors are excluded, thus generating filtered short-cycle trigger factors and long-cycle trigger factors.

[0024] Calculate the continuous change in the offset difference between the filtered short-period triggering factor and the long-period triggering factor, and accumulate the continuous change over multiple consecutive time periods to form an offset difference accumulation sequence.

[0025] A dual-condition determination is made based on the cumulative amount and duration in the cumulative offset sequence. When the cumulative amount exceeds a preset associated offset threshold and the duration exceeds a preset first duration threshold, a behavioral mutation event is generated. The behavioral mutation event includes a mutation category to indicate the mutation type, a mutation amplitude to indicate the mutation intensity, and a mutation time period to indicate the mutation occurrence range.

[0026] Preferably, the strategy migration process is performed based on the behavioral mutation event to generate a migrated waste disposal management strategy, specifically including an updated waste bin overflow prediction window, an adjusted waste collection priority, and a reset waste disposal classification correction prompt, including:

[0027] The prediction window for overflowing trash cans is dynamically expanded or contracted based on the magnitude and type of the behavioral mutation event, so that the time range of the prediction window changes proportionally with the magnitude of the mutation, and an adjusted prediction window is generated.

[0028] Based on the waste categories involved in behavioral mutation events and their frequency of occurrence, the collection objects are prioritized. By establishing a collection priority enhancement factor, high mutation categories are given higher collection priority, and an adjusted collection order is generated.

[0029] Based on the waste sorting deviations corresponding to behavioral mutation events, the deviation category is used as a key prompt category. The frequency of push notifications for this category is increased, and the prompt content is adaptively updated to generate updated waste disposal and sorting correction prompts.

[0030] Preferably, the short-cycle volatility indicator and the long-cycle stability indicator are time-aligned, and the trend difference value between the two is calculated within the common time period after alignment. A preliminary correlation result is then generated based on the trend difference value, including:

[0031] Time reference points are determined based on short-cycle fluctuation indicators and long-cycle stability indicators, and these time reference points are expanded to form a reference time series that describes the time distribution of the two types of indicators.

[0032] Based on the reference time series, short-cycle volatility indicators and long-cycle stability indicators are time-mapped so that the two types of indicators correspond on the same time axis, forming an aligned indicator sequence.

[0033] Based on the alignment indicator sequence, the direction, magnitude, and speed of change of short-cycle volatility indicators and long-cycle stability indicators within the same time period are obtained, and interval differences are extracted to generate trend difference values.

[0034] Based on the trend difference value, the deviation direction, deviation intensity and deviation duration within the interval are correlated to obtain the primary correlation results used to characterize the relationship of cross-cycle behavior change.

[0035] Preferably, based on the trend difference values ​​in the primary correlation results, a segmented analysis is performed on the directionality, cumulativeity, and persistence of the differences within the alignment time period to generate basic values ​​for short-term triggering factors describing short-term burst behavior and basic values ​​for long-term triggering factors describing long-term offset behavior, including:

[0036] The trend difference values ​​are divided directionally according to the direction of change of the trend difference values, forming directional difference segments to represent the direction of trend change;

[0037] Based on the directional difference segment, the trend difference values ​​within the difference segment are accumulated, and a cumulative difference segment is formed to represent the intensity of the cumulative difference based on the characteristics of the cumulative change.

[0038] Based on the duration of the cumulative difference segment within the alignment time period, a persistence analysis is performed on the difference segment to identify the difference segment whose duration exceeds a preset second persistence threshold, and a short-cycle trigger factor base value is generated based on the difference segment.

[0039] Long-term offset analysis is performed on the difference segments that do not meet the continuous threshold. Trend aggregation is performed based on the offset direction, offset magnitude and offset stage characteristics to generate the base value of long-term trigger factor.

[0040] Secondly, a waste disposal optimization management system based on classification data mining, the system comprising:

[0041] The data acquisition module is used to acquire waste disposal behavior data uploaded by the waste disposal collection device, and generate an original behavior sequence to represent continuous disposal behavior based on the data. The waste disposal behavior data includes waste disposal amount, waste disposal time and waste disposal type.

[0042] The sequence construction module is used to divide the original behavior sequence into time periods, generate short-cycle behavior data and long-cycle behavior data, and construct short-cycle classification sequences and long-cycle classification sequences respectively.

[0043] The indicator generation module is used to perform category difference processing between adjacent time periods and category aggregation calculation for each long period based on the short-period classification sequence and the long-period classification sequence, to generate short-period volatility indicators and long-period stability indicators.

[0044] The correlation analysis module is used to perform cross-series correlation analysis based on short-cycle volatility indicators and long-cycle stability indicators, generate correlation analysis results, and construct hierarchical triggering factors including short-cycle triggering factors and long-cycle triggering factors based on them.

[0045] The mutation identification module is used to determine the offset difference based on the hierarchical triggering factor. When the offset difference between the short-cycle triggering factor and the long-cycle triggering factor exceeds the preset associated offset threshold, a behavioral mutation event is generated.

[0046] The strategy migration module is used to perform strategy migration processing based on behavioral mutation events, and generate the migrated waste disposal management strategy, which includes an updated waste bin overflow prediction window, an adjusted waste collection priority, and a reset waste disposal classification correction prompt.

[0047] The waste disposal management module is used to perform waste disposal optimization management operations according to the migrated waste disposal management strategy, and output the execution results to the waste disposal management terminal.

[0048] The above-described solution of the present invention has at least the following beneficial effects:

[0049] First, by continuously processing the multi-dimensional waste disposal behavior data uploaded by waste collection devices, this invention can reconstruct fragmented single disposal records into complete original behavioral sequences, freeing behavioral patterns from the constraints of fixed-time average statistics. This serialized representation avoids the problem of missing behavioral patterns caused by data breakpoints in traditional rule-based models, enabling subsequent analysis to directly reflect changes in residents' actual waste disposal rhythm and providing a continuous and stable data foundation for further identification of behavioral trends and anomalies.

[0050] Building upon this foundation, this invention divides the original behavioral sequence into two time scales: short-term and long-term. It then constructs short-term and long-term classification sequences respectively, enabling behavioral features to be expressed at different temporal granularities. Short-term sequences can sensitively capture rapid fluctuations in delivery volume or category within short time periods, while long-term sequences are used to present the dominant changing trends at different stages. Unlike existing technologies that rely on a single time granularity to judge behavior, the periodic hierarchical structure of this invention effectively improves the model's ability to distinguish between sudden behaviors and long-term changes, reducing analytical distortions and misjudgments that occur in scenarios involving changes in user groups, as is common with traditional methods.

[0051] Furthermore, this invention generates short-cycle fluctuation indicators and long-cycle stability indicators by performing category differencing on short-cycle classification sequences and aggregation calculations on long-cycle classification sequences, respectively, thus quantifying behavioral changes in two dimensions: "rapid fluctuation" and "overall stability." Compared to existing technologies that only use simple thresholds or clustering results as judgment criteria, this invention can extract more layers of information from the behavioral structure itself, enabling the system to have higher sensitivity and analytical capabilities when facing changes in residents' delivery habits.

[0052] Building upon this, the present invention constructs hierarchical triggering factors through cross-sequence association analysis, transforming the discrepancies between short-term volatility indicators and long-term stability indicators into triggering factors with hierarchical attributes. Short-term triggering factors express the impact of short-term sudden behaviors, while long-term triggering factors describe the degree to which behaviors deviate from long-term trends. Unlike existing technologies that cannot distinguish between short-term anomalies and long-term structured changes, the hierarchical triggering factors of this invention enable the system to make more refined judgments based on the nature of behavioral changes, significantly reducing scheduling errors caused by misinterpreting short-term fluctuations as trend changes.

[0053] Furthermore, by determining the offset difference between the stratified triggering factors, when a difference exceeding a preset correlation offset threshold occurs between the short-cycle triggering factor and the long-cycle triggering factor, this invention can generate a behavioral mutation event to clearly identify that the behavioral pattern has deviated from the original trend. This mechanism enables the system to identify structural changes in residents' waste disposal behavior in advance, avoiding situations like the "early overflow caused by concentrated waste disposal in the early morning" described in the background art, which are not recognized by the system, effectively reducing waste overflow and environmental impact caused by scheduling delays.

[0054] Finally, by implementing strategy migration processing based on behavioral mutation events, this invention can automatically update the garbage bin overflow prediction window, garbage collection priority, and garbage disposal classification correction prompts, achieving dynamic adjustment of management strategies. This strategy migration mechanism allows the system scheduling results to evolve in real time with changes in disposal behavior, rather than relying on long-term fixed rules, thus improving prediction accuracy, collection scheduling adaptability, and the targeted nature of classification guidance. In actual community operation, for example, when a certain type of cardboard waste suddenly increases and significantly deviates from long-term patterns within a short period, this invention can promptly generate mutation events and automatically increase the collection priority of the relevant garbage bins, preventing overflow in the early morning, reducing the frequency of manual intervention by management personnel, and maintaining the stability and cleanliness of the community environment. Attached Figure Description

[0055] Figure 1 This is a flowchart of a waste disposal optimization management method based on classification data mining provided in an embodiment of the present invention. Detailed Implementation

[0056] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0057] like Figure 1 As shown, embodiments of the present invention propose a waste disposal optimization management method based on classification data mining, the method comprising:

[0058] The waste disposal behavior data uploaded by the waste disposal collection device is acquired, and an original behavior sequence representing continuous disposal behavior is generated based on the data. The waste disposal behavior data includes the amount of waste disposed of, the time of waste disposal, and the type of waste disposal.

[0059] The original behavior sequence is divided into time periods to generate short-cycle behavior data and long-cycle behavior data, and short-cycle classification sequences and long-cycle classification sequences are constructed respectively.

[0060] Based on the short-cycle classification sequence and the long-cycle classification sequence, the category difference processing of adjacent time periods and the category aggregation calculation of each long-cycle stage are performed to generate short-cycle fluctuation index and long-cycle stability index.

[0061] Cross-sequence correlation analysis is performed based on short-cycle volatility indicators and long-cycle stability indicators to generate correlation analysis results, and a hierarchical triggering factor including short-cycle triggering factors and long-cycle triggering factors is constructed based on these results.

[0062] The offset difference is determined based on the hierarchical triggering factor. When the offset difference between the short-cycle triggering factor and the long-cycle triggering factor exceeds the preset associated offset threshold, a behavioral mutation event is generated.

[0063] Based on the behavioral mutation event, the strategy migration process is executed to generate the migrated waste disposal management strategy, which includes an updated waste bin overflow prediction window, an adjusted waste collection priority, and a reset waste disposal classification correction prompt.

[0064] Execute waste disposal optimization management operations according to the migrated waste disposal management strategy, and output the execution results to the waste disposal management terminal.

[0065] In this embodiment of the invention, by continuously processing the multi-dimensional waste disposal behavior data uploaded by the waste disposal collection device, it is possible to obtain an original behavior sequence that is more consistent with the actual changes in disposal behavior, avoiding the problem of incomplete behavior patterns caused by directly using scattered data, thereby providing a complete and continuous data foundation for subsequent behavior analysis.

[0066] By dividing the original behavioral sequence into time periods and constructing short-term and long-term classification sequences respectively, the delivery behavior can be characterized at different time scales. On the one hand, the short-term classification sequence can reflect the fluctuations of delivery behavior within a smaller time range; on the other hand, the long-term classification sequence can reflect the trend changes of delivery behavior within a larger time range. This hierarchical sequence structure can improve the ability to analyze the characteristics of delivery behavior, effectively reflecting both short-term fluctuations and long-term changes in delivery behavior.

[0067] By performing category differencing on short-term classification sequences and clustering calculations on long-term classification sequences, short-term volatility indicators and long-term stability indicators can be obtained, respectively. Short-term volatility indicators reflect the magnitude of changes within a short period, revealing whether rapid changes occur in campaign behavior; long-term stability indicators reflect the concentration of categories at different stages, demonstrating the stable trend of campaign behavior over a longer period. These two types of indicators complement each other, facilitating the construction of a more complete behavioral characteristic system.

[0068] By utilizing short-cycle volatility indicators and long-cycle stability indicators for cross-series correlation analysis, the relationship between behavioral changes across two time scales can be identified, thereby constructing hierarchical triggering factors. Short-cycle triggering factors express the impact of short-term fluctuations, while long-cycle triggering factors express the deviation effect of long-term trends. The construction of hierarchical triggering factors enables the system to identify potential sudden changes or trend deviations in deployment behavior from multiple levels.

[0069] By analyzing the offset difference between tiered triggering factors, a behavioral mutation event can be generated when the offset difference between short-cycle and long-cycle triggering factors exceeds a set offset threshold. Behavioral mutation events reflect significant deviations between the deployment behavior and the expected pattern, providing a triggering basis for subsequent strategy adjustments.

[0070] By implementing strategy migration processing based on behavioral mutation events, overflow prediction, collection priority, and disposal classification prompts in waste disposal management strategies can be updated, allowing management strategies to dynamically adjust as behavior changes. This dynamic migration approach ensures that management strategies align with actual disposal behavior, improving the accuracy of disposal guidance, waste collection, and disposal scheduling.

[0071] For example, in the daily operation of a residential area, waste disposal behavior data is collected to form a raw behavior sequence, which is then divided into short-term and long-term classification sequences based on time periods. If there is a sudden increase in the amount of a certain type of waste disposed of within the short-term sequence, while this type of waste does not remain stable in the long-term classification sequence, then a significant difference will emerge between the short-term fluctuation index and the long-term stability index during the correlation analysis phase. This difference, amplified by a stratified triggering factor, may exceed the offset threshold, thus generating a behavioral mutation event, indicating a change in the disposal pattern of this type of waste. The subsequently generated migration management strategy will adjust the collection priority of this type of waste and update the corresponding disposal prompts, so that the collection and transportation arrangements are more in line with the current residents' disposal behavior.

[0072] In a preferred embodiment of the present invention, waste disposal behavior data uploaded by the waste disposal collection device is acquired, and an original behavior sequence representing continuous disposal behavior is generated based on the data, specifically including:

[0073] In the acquisition step, the data uploaded by the collection devices deployed at the garbage disposal points are received in chronological order, and the garbage disposal amount, garbage disposal time and garbage disposal type in each data record are structured so that the data can be stored in a unified format.

[0074] The structured data is sorted according to the delivery time to ensure that there is no time overlap or interruption between data, thereby ensuring the continuity of behavioral data in the time dimension.

[0075] Based on the sorted data, the amount of waste disposed of, the type of waste disposed of, and the time of disposal at different time points are integrated to form a continuously arranged dataset. This dataset is used to represent the temporal evolution of waste disposal behavior and serves as the original behavior sequence for subsequent cycle construction and behavior change analysis.

[0076] When the time interval between two adjacent data exceeds the set maximum allowable interval, blank data or zero-value data can be added in the middle to keep the behavior sequence logically continuous, thereby avoiding the time discontinuity problem caused by sparse delivery.

[0077] In a preferred embodiment of the present invention, the original behavioral sequence is divided into time periods to generate short-cycle behavioral data and long-cycle behavioral data, and short-cycle classification sequences and long-cycle classification sequences are constructed respectively, specifically including:

[0078] By reading the delivery time corresponding to each record in the original behavior sequence, the complete time range is divided into multiple time periods. The length of short-cycle time periods can be selected at the hour level or other relatively short time units; while long-cycle time periods are determined based on the time span of days, weeks or months.

[0079] The data in the original behavior sequence are aggregated according to the short-cycle period and the long-cycle period, so that the data in each short-cycle period and the long-cycle period are integrated into the corresponding short-cycle behavior data and long-cycle behavior data.

[0080] Based on the quantity and proportion of waste disposal types appearing in each time period in the short-term behavioral data, they are summarized into preset classification categories to form a short-term classification sequence, which is used to describe behavioral changes within a small time range.

[0081] Based on the main delivery types and their changing trends in each stage of long-term behavioral data, the behavioral data is aggregated into long-term classification results to form a long-term classification sequence, which is used to describe the overall trend and structural changes.

[0082] Among them, short-cycle classification sequences can more sensitively reflect short-term fluctuations in deployment, while long-cycle classification sequences can reflect long-term behavioral patterns. The combination of the two can provide a multi-dimensional data foundation for the generation of subsequent indicators.

[0083] In a preferred embodiment of the present invention, a preset correlation offset threshold is used to determine whether the offset difference between the short-cycle triggering factor and the long-cycle triggering factor is sufficient to indicate a behavioral mutation, specifically including:

[0084] Based on historical waste disposal behavior data, the difference range between short-cycle triggering factors and long-cycle triggering factors under normal conditions is statistically analyzed to obtain the offset difference range under multiple normal behavior patterns.

[0085] The above-mentioned offset ranges are organized, and based on the frequency of offset occurrence, duration and stable range in normal behavior, the upper limit value that the offset can remain stable under normal behavior is determined as the basis for setting the offset threshold.

[0086] By analyzing historical abnormal cases or simulated abnormal scenarios, we can observe the changes in triggering factors caused by abnormal behavior, compare the deviation characteristics that appear in abnormal behavior with the normal range, and determine the minimum deviation amplitude that can characterize abnormal behavior.

[0087] By combining the upper limit of normal offset difference and the lower limit of abnormal offset difference, the difference range between the two is used as an adjustable space, and an appropriate value is selected as the preset associated offset threshold according to the application scenario.

[0088] This threshold is not only used to detect the magnitude of the offset, but can also be combined with the setting rules for the offset duration to trigger behavioral mutation events, making the judgment of behavioral mutations more consistent with the actual deployment changes.

[0089] In a preferred embodiment of the present invention, based on the short-period classification sequence and the long-period classification sequence, category difference processing of adjacent time periods and category aggregation calculation of each long-period stage are performed respectively to generate short-period fluctuation indicators and long-period stability indicators, including:

[0090] The short-period classification sequence is segmented according to a fixed-length short-period analysis window. Within each analysis window, the directional comparison, continuity judgment and amplitude accumulation of the category changes in adjacent time periods are performed to generate short-period difference results within the short-period window.

[0091] Consistency detection is performed on the category change patterns of multiple consecutive analysis windows based on the short-cycle difference results, and the consistency detection results are transformed into short-cycle fluctuation indicators.

[0092] Long-period classification sequences are grouped according to long-period stages. Within each long-period group, long-period clustering results are generated based on the frequency of category occurrence, duration, and stability of the dominant category within the stage. Long-period stability indices are then calculated based on the long-period clustering results.

[0093] In this embodiment of the invention, short-cycle classification sequences are segmented according to a fixed-length analysis window, enabling short-cycle behavior to be analyzed in a continuous and structured manner. By comparing the directionality, judging the continuity, and accumulating the amplitude of category changes in adjacent time periods, the change features within a short period can be integrated into short-cycle difference results, avoiding the fluctuation interference caused by single-point data and helping to form a more reliable short-cycle feature representation.

[0094] Consistency testing across multiple consecutive analysis windows can identify persistent trends within short-term fluctuations, providing greater stability than differential results from a single window. The resulting short-term volatility index reflects the overall degree of behavioral change within short periods, thus offering a more accurate description of rapid fluctuations.

[0095] Grouping long-term classification sequences according to long-term stages allows the classification results to reflect a larger time span. Cluster analysis of category frequency, duration, and the stability of the dominant category provides a concentrated expression of long-term behavioral trends. The resulting long-term stability index reflects the dominant patterns of campaign behavior across different long-term stages, providing a clearer basis for identifying subsequent trend changes.

[0096] This processing method can obtain characteristic indicators representing behavioral fluctuations and behavioral stability in both short-term and long-term time dimensions, making the subsequent correlation analysis reveal the relationship of behavioral changes more fully.

[0097] In a preferred embodiment of the present invention, the short-period classification sequence is segmented according to a fixed-length short-period analysis window. Within each analysis window, the directional comparison, continuity judgment, and amplitude accumulation of category changes in adjacent time periods are performed to generate short-period difference results within the short-period window, specifically including:

[0098] The short-period classification sequence is segmented according to the set short-period analysis window length, so that each analysis window corresponds to a fixed time range.

[0099] Within each analysis window, the classification results of adjacent time periods are read in chronological order, and the direction of category change between the two is compared to determine whether the category change is an increase, a decrease, or no change.

[0100] Based on the continuity of category changes between adjacent time periods, determine whether the behavior shows a consistent trend of change within a short period of time, such as continuous rise, continuous fall, or unstable trend.

[0101] Based on the magnitude of the category change, the changes between multiple adjacent time periods are accumulated so that the total magnitude of change within the analysis window can reflect the overall fluctuation.

[0102] Based on the combined characteristics of directionality, continuity, and cumulative amplitude, the overall changes within the analysis window are transformed into short-period difference results, which are used to represent the comprehensive characteristics of behavioral changes within the short-period window.

[0103] In a preferred embodiment of the present invention, consistency detection is performed on the category change patterns of multiple consecutive analysis windows based on the short-period difference results, and the consistency detection results are converted into short-period fluctuation indicators, specifically including:

[0104] Arrange the short-cycle difference results generated in multiple adjacent short-cycle analysis windows in chronological order and check the continuity of their changing trends;

[0105] In multiple consecutive analysis windows, determine whether the difference results show a consistent direction of change, such as whether they form a pattern of continuous increase, continuous decrease, or continuous instability.

[0106] Based on the consistency judgment results, extract the behavioral pattern characteristics between continuous windows, including whether there are short-term concentrated changes, short-term periodic fluctuations or short-term abnormal behaviors.

[0107] The above behavioral pattern features are summarized and processed according to preset classification rules, so that the behavioral patterns can be expressed as quantifiable fluctuation features.

[0108] Based on the strength, frequency, and duration of the fluctuation characteristics, the consistency detection results are converted into short-cycle fluctuation indicators to represent the overall fluctuation level of behavioral changes within a short time range.

[0109] In a preferred embodiment of the present invention, the long-period classification sequence is grouped according to long-period stages. Within each long-period group, a long-period clustering result is generated based on the frequency of category occurrence, duration, and stability of the dominant category within the stage. A long-period stability index is then calculated based on the long-period clustering result, specifically including:

[0110] The long-cycle classification sequence is divided according to a preset long-cycle stage, so that each stage can cover a sufficiently long time range to reflect long-term change patterns.

[0111] Within each long-term phase, the occurrence of different categories is statistically analyzed, and the frequency of occurrence and duration of each type of waste in that phase are recorded.

[0112] Based on the statistical frequency and duration, the dominant category is determined according to the degree of dominance of different categories within the period, and the stability of the dominant category is further determined, that is, whether the dominant category continues to maintain its dominant position within the period.

[0113] Based on the combined performance of frequency, duration, and stability of the dominant category, the classification features of this stage are used to generate long-term clustering results to represent the main behavioral features of this long-term stage.

[0114] Based on the aggregation results, the stability of each long-term stage is summarized and analyzed to form a long-term stability index, which can reflect whether long-term behavior remains stable or changes in trend.

[0115] In a preferred embodiment of the present invention, cross-series correlation analysis is performed based on short-cycle volatility indicators and long-cycle stability indicators to generate correlation analysis results, and a hierarchical triggering factor including short-cycle triggering factors and long-cycle triggering factors is constructed based on these results, including:

[0116] The short-cycle volatility indicator and the long-cycle stability indicator are time-aligned, and the trend difference value between the two is calculated within the common time period after alignment. The initial correlation result is generated based on the trend difference value.

[0117] Based on the trend difference values ​​in the primary correlation results, the directionality, cumulativeity, and persistence of the difference in the alignment time period are segmented for analysis, generating basic values ​​of short-term triggering factors to describe short-term burst behavior and basic values ​​of long-term triggering factors to describe long-term offset behavior.

[0118] Based on the structural differences between the base values ​​of short-cycle trigger factors and long-cycle trigger factors, the two are hierarchically reorganized, with the short-cycle trigger factor placed in the rapid response layer and the long-cycle trigger factor placed in the trend identification layer, forming a layered trigger factor with two layers.

[0119] In this embodiment of the invention, by aligning the short-cycle volatility indicators and the long-cycle stability indicators over time, the two types of indicators can be compared within the same time period, avoiding analytical bias caused by differences in time scales. The time-aligned indicators can accurately reflect the relationship between short-term changes and long-term trends within the same time frame.

[0120] By calculating trend difference values ​​within the aligned time period, the degree of difference between short-term fluctuations and long-term stability can be obtained, including differences in the direction, magnitude, and rate of change. This difference analysis can reveal whether investment behavior deviates from short-term fluctuations and long-term trends, laying the foundation for further construction of triggering factors.

[0121] Segmented analysis of trend variance values ​​allows for the differentiation and separate expression of the directionality, cumulativeity, and persistence of these variances. Directional analysis identifies the direction of change in variances, cumulative analysis identifies whether variances are accumulating, and persistence analysis identifies whether variances persist for a certain duration. Further analysis of these variance characteristics generates basic values ​​for short-term and long-term triggering factors, enabling the independent expression of behavioral changes across different time dimensions.

[0122] By hierarchically reorganizing the basic values ​​of short-cycle triggering factors and long-cycle triggering factors, a hierarchical triggering factor structure with different reaction speeds and different behavioral characteristic focuses can be formed, thereby providing a more comprehensive basis for subsequent identification of behavioral mutations.

[0123] In a preferred embodiment of the present invention, based on the structural differences between the base values ​​of short-cycle triggering factors and long-cycle triggering factors, the two are hierarchically reorganized. The short-cycle triggering factor is placed in the rapid response layer, and the long-cycle triggering factor is placed in the trend identification layer, forming a hierarchical triggering factor containing two layers, specifically including:

[0124] The basic values ​​of short-cycle triggering factors and long-cycle triggering factors are classified according to their characteristic sources. The basic values ​​of short-cycle triggering factors come from short-cycle sudden behavior analysis, while the basic values ​​of long-cycle triggering factors come from long-term deviation behavior analysis.

[0125] Based on the characteristics of sudden and rapid behavioral changes in the basic value of short-cycle triggering factors, they are classified into the rapid response layer, which enables this layer to reflect behavioral changes that occur in a short period of time in a timely manner.

[0126] Based on the characteristics of strong cumulative change and long-term deviation trend of behavior in the basic value of long-cycle triggering factors, it is classified into the trend identification layer, so that this layer can be used to represent the degree of deviation of behavior changes over a longer period.

[0127] Based on the different functions of the two levels, the structural relationship between the two types of triggering factors is combined to form a top-down hierarchical arrangement between the rapid response layer and the trend recognition layer, thereby generating a layered triggering factor with a two-layer structure.

[0128] This hierarchical triggering factor can simultaneously reflect the sudden changes in short-cycle behavior and the offset characteristics of long-cycle behavior, providing a more comprehensive information basis for subsequent offset difference determination.

[0129] In a preferred embodiment of the present invention, a preset second persistence threshold is used to determine whether the offset difference maintains a stable trend over multiple consecutive time periods, specifically including:

[0130] Based on the continuous characteristics of short-term and long-term changes in historical waste disposal behavior data, the duration distribution range of the deviation under normal behavior conditions is statistically analyzed to identify common duration intervals under normal conditions.

[0131] After analyzing historical data, it was determined that deviations occurring over shorter periods of normal behavior were mostly sporadic changes, while deviations occurring over longer periods were more likely to reflect changes in behavioral trends.

[0132] Based on the above analysis, by comparing the maximum duration of normal behavior with the minimum duration of abnormal behavior, a duration boundary range that can distinguish between normal short-term fluctuations and long-term trend deviations can be formed.

[0133] Taking into account management scenarios, such as the frequency of garbage disposal in residential areas and the distribution of peak disposal times each day, a time value that can reliably reflect the persistence of behavioral changes is selected as the preset second persistence threshold within the above-mentioned boundary range.

[0134] When the cumulative offset sequence remains above the preset second persistence threshold for multiple consecutive time periods, the behavioral change can be considered to have a persistent characteristic, thus determining that the behavioral mutation event has occurred.

[0135] In a preferred embodiment of the present invention, offset difference is determined based on hierarchical triggering factors. When the offset difference between the short-cycle triggering factor and the long-cycle triggering factor exceeds a preset associated offset threshold, a behavioral mutation event is generated, including:

[0136] Offset source filtering is performed on the short-cycle trigger factors and long-cycle trigger factors of the layered trigger factors respectively. The instantaneous changes caused by a single anomaly in the short-cycle trigger factors are removed, and the unstable offsets caused by short-term fluctuations in the long-cycle trigger factors are excluded, thus generating filtered short-cycle trigger factors and long-cycle trigger factors.

[0137] Calculate the continuous change in the offset difference between the filtered short-period triggering factor and the long-period triggering factor, and accumulate the continuous change over multiple consecutive time periods to form an offset difference accumulation sequence.

[0138] A dual-condition determination is made based on the cumulative amount and duration in the cumulative offset sequence. When the cumulative amount exceeds a preset associated offset threshold and the duration exceeds a preset first duration threshold, a behavioral mutation event is generated. The behavioral mutation event includes a mutation category to indicate the mutation type, a mutation amplitude to indicate the mutation intensity, and a mutation time period to indicate the mutation occurrence range.

[0139] In this embodiment of the invention, by performing offset source filtering on short-cycle triggering factors and long-cycle triggering factors respectively, the instantaneous changes caused by a single abnormal delivery in short-cycle triggering factors can be eliminated, and the unstable offsets caused by short-term disturbances in long-cycle triggering factors can be eliminated, so that the filtered triggering factors can better represent the true trend of behavior changes and reduce the risk of misjudgment.

[0140] By calculating the continuous change in the offset difference between the filtered short-cycle triggering factors and the long-cycle triggering factors, and forming an offset difference cumulative sequence, the changes in the offset difference over multiple consecutive time periods can be reflected, simultaneously showing the persistence and cumulative nature of behavioral changes. Compared to single-point offset values, the cumulative sequence can provide more comprehensive behavioral offset information.

[0141] By using a dual-condition judgment based on the cumulative amount and duration of the offset accumulation sequence, a true behavioral mutation can be effectively identified. A behavioral mutation event is only generated when the cumulative amount of the offset reaches a set threshold and the duration exceeds a set range, making the judgment conditions for behavioral mutations more reliable.

[0142] The generated behavioral mutation events include mutation type, mutation magnitude, and mutation time period, which can clearly describe the nature, intensity, and scope of behavioral changes, providing a sufficient information basis for subsequent policy migration processing.

[0143] In a preferred embodiment of the present invention, the short-cycle triggering factors and long-cycle triggering factors of the hierarchical triggering factors are respectively subjected to offset source filtering. Instantaneous changes caused by single anomalies in the short-cycle triggering factors are removed, and unstable offsets caused by short-term fluctuations in the long-cycle triggering factors are excluded, generating filtered short-cycle triggering factors and long-cycle triggering factors. Specifically, this includes:

[0144] Obtain the changes of short-cycle triggering factors over multiple consecutive time periods, and identify sudden increases or decreases in triggering factors within a single time period;

[0145] Based on the persistence of changes in short-cycle triggering factors within adjacent time periods, it can be determined whether a sudden increase or decrease occurs only in a single time period. When a change only occurs in an isolated time period and does not persist in the time periods before or after it, this type of change is considered an instantaneous change.

[0146] The above-mentioned instantaneous changes are removed from the short-cycle triggering factors so that the short-cycle triggering factors can reflect real and continuous behavioral fluctuations within a short time range.

[0147] Obtain the changing trend of long-term triggering factors over a long period of time, and identify the small offsets caused by short-term disturbances;

[0148] Based on the phased change characteristics of long-cycle triggering factors, determine whether the offset only occurs in a local time period and has not formed a trend change. When the offset is only a short-term disturbance, it is excluded from the long-cycle triggering factors.

[0149] Through the above processing, filtered short-period triggering factors and long-period triggering factors are generated, so that both retain their corresponding behavioral characteristics and avoid the subsequent offset difference calculation being affected by noise.

[0150] In a preferred embodiment of the present invention, the continuous change in the offset difference between the filtered short-period triggering factor and the long-period triggering factor is calculated, and the continuous change is accumulated over multiple consecutive time periods to form an offset difference accumulation sequence, specifically including:

[0151] Obtain the values ​​of the filtered short-cycle trigger factor and long-cycle trigger factor in each time period, and calculate the difference between the two in the same time period to obtain the offset difference in that time period.

[0152] Based on the changes in the offset over adjacent time periods, determine whether the offset shows a trend of continuous increase, continuous decrease, or stability.

[0153] The changes in the offset over multiple consecutive time periods are accumulated so that the cumulative result can reflect the overall degree of change of the offset over time, avoiding bias in behavior judgment caused by a single time period.

[0154] When the offset difference shows a consistent direction and a continuous change over a long period of time, these continuous changes are accumulated to form an offset difference accumulation sequence, which can reflect the cumulative trend of the offset difference over a long period of time.

[0155] The cumulative offset sequence is used for subsequent persistence determination, enabling the identification of behavioral mutation events to consider both the offset magnitude and the offset duration.

[0156] In a preferred embodiment of the present invention, a dual-condition determination is performed based on the accumulated amount and duration in the offset accumulation sequence. When the accumulated amount exceeds a preset associated offset threshold and the duration exceeds a preset first persistence threshold, a behavioral mutation event is generated. The behavioral mutation event includes a mutation category representing the mutation type, a mutation amplitude representing the mutation intensity, and a mutation time period representing the mutation occurrence range, specifically including:

[0157] Analyze the cumulative offset sequence to identify the time period when the cumulative amount reaches its peak, and record the magnitude of the cumulative amount;

[0158] Based on the cumulative trend of the cumulative offset sequence, it is determined whether the cumulative amount exceeds the preset associated offset threshold, so that the judgment result can reflect whether the offset has reached the degree of change that needs attention.

[0159] The time periods in which consecutive offsets occur in the cumulative offset sequence are statistically analyzed to identify whether their duration exceeds a preset first duration threshold.

[0160] When the accumulated amount exceeds the preset associated offset threshold and the duration exceeds the preset first duration threshold, the situation where both conditions are met simultaneously is identified as a behavioral mutation event.

[0161] Based on the direction and category of the offset changes in the cumulative offset sequence, the mutation categories involved in the behavioral mutation events are labeled;

[0162] Based on the magnitude of the accumulated amount in the offset accumulation sequence, it is converted into a mutation magnitude to represent the intensity of the mutation event;

[0163] The mutation time period is determined by the start and end times of continuous offsets, so that the range of mutation events can be clearly expressed.

[0164] The resulting behavioral mutation events can serve as the basis for strategy migration processing, providing a clear data source for subsequent adjustments to waste disposal management strategies.

[0165] In a preferred embodiment of the present invention, a strategy migration process is performed based on the behavioral mutation event to generate a migrated waste disposal management strategy. Specifically, this includes an updated waste bin overflow prediction window, an adjusted waste collection priority, and a reset waste disposal classification correction prompt, including:

[0166] The prediction window for overflowing trash cans is dynamically expanded or contracted based on the magnitude and type of the behavioral mutation event, so that the time range of the prediction window changes proportionally with the magnitude of the mutation, and an adjusted prediction window is generated.

[0167] Based on the waste categories involved in behavioral mutation events and their frequency of occurrence, the collection objects are prioritized. By establishing a collection priority enhancement factor, high mutation categories are given higher collection priority, and an adjusted collection order is generated.

[0168] Based on the waste sorting deviations corresponding to behavioral mutation events, the deviation category is used as a key prompt category. The frequency of push notifications for this category is increased, and the prompt content is adaptively updated to generate updated waste disposal and sorting correction prompts.

[0169] In this embodiment of the invention, the prediction window for garbage bin overflow is dynamically expanded or contracted based on behavioral mutation events, allowing the prediction window to adjust with changes in behavior. When a mutation event reflects a significant increase in a certain type of garbage within a specific time period, expanding the prediction window can anticipate potential overflow risks; when the behavioral trend shifts towards reduction, contracting the prediction window can reduce unnecessary prediction redundancy. This makes the prediction results closer to the actual disposal schedule and improves the timeliness of prediction adjustments.

[0170] By prioritizing waste collection targets based on the type and frequency of occurrence in behavioral mutation events, the collection order can be automatically rearranged as behavior changes. When the disposal behavior of a certain type of waste shows a high frequency or significant increase during a mutation event, raising the collection priority of this type of waste helps reduce the risk of accumulation and makes the collection schedule more consistent with the current disposal status. Priority enhancement factors further clarify the priority relationships between different categories, making the collection task arrangement clearer.

[0171] By updating the prompts based on waste sorting deviations during behavioral mutation events, the information can be adjusted to reflect actual changes in waste disposal behavior. When a certain type of waste exhibits a trend of sorting deviation during a mutation event, increasing the frequency of prompts for that category and updating the prompt content helps guide disposal behavior back to correct sorting methods, reducing the recurrence of disposal errors. Overall, this method enables waste disposal guidance, collection scheduling, and overflow prediction to maintain linkage, improving the coordination of the waste management process.

[0172] In a preferred embodiment of the present invention, the prediction window for a overflowing trash can is dynamically expanded or contracted based on the magnitude and category of the behavioral mutation event, so that the time range of the prediction window changes proportionally with the magnitude of the mutation, and an adjusted prediction window is generated. Specifically, this includes:

[0173] Identify the mutation categories contained in behavioral mutation events and determine whether the mutation category belongs to the type of waste with a high risk of overflowing the trash can, such as wet waste or large-volume waste that takes up a lot of space;

[0174] Based on the magnitude of the mutation recorded in the behavioral mutation event, the degree of change in the delivery volume can be judged. When the mutation magnitude is large, it indicates that a higher delivery load may be generated in a short period of time.

[0175] The magnitude of the mutation is matched with the preset window adjustment rules. For example, when the magnitude of the mutation is high, the prediction window is expanded forward appropriately so that the system can identify the overflow trend earlier; when the magnitude of the mutation is low, the prediction window is shrunk accordingly based on the characteristics of the mutation category to avoid the prediction range being too large and causing the prediction results to be biased.

[0176] During the adjustment of the prediction window, the adjustment ratio is determined based on the characteristics of the disposal behavior of the mutation category. For example, for waste categories with high disposal concentration, the window expansion ratio can be appropriately increased.

[0177] Based on the above adjustments, a new forecast window time range is formed to more accurately reflect the impact of sudden events on the overflow trend.

[0178] The final output is the adjusted prediction window, which serves as the basis for subsequent overflow prediction processes.

[0179] In a preferred embodiment of the present invention, the waste collection objects are prioritized according to the waste categories involved in the behavioral mutation events and their frequency of occurrence. By establishing a collection priority enhancement factor, high mutation categories are given higher collection priority, and an adjusted collection order is generated, specifically including:

[0180] Analyze the types of waste involved in behavioral mutation events, identify which types of waste exhibit more obvious mutation behaviors, and record their frequency of occurrence within a preset time period;

[0181] Based on the combination of frequency and magnitude of mutations, the urgency of mutation behavior is determined. For example, categories with high mutation magnitude and high frequency have a more significant impact on waste disposal needs.

[0182] Based on the importance of the mutation behavior, a collection priority enhancement factor is constructed to improve the priority of the mutation category; this enhancement factor can be set according to the capacity occupancy characteristics of the waste category, the usage scenario, and the management rules.

[0183] The waste collection targets are sorted according to their priority after adjustment by the boost factor, so that waste bins of the mutation category can be prioritized for collection.

[0184] Based on the sorting results, a final collection order is generated, and this order is used to generate subsequent scheduling instructions to improve the rationality of the collection arrangement.

[0185] In a preferred embodiment of the present invention, based on the waste sorting deviation corresponding to the behavioral mutation event, the deviation category is used as a key prompt category, the push frequency of prompts for this category is increased, and the prompt content is adaptively updated to generate an updated waste disposal and sorting correction prompt, specifically including:

[0186] Analyze the mutation categories recorded in behavioral mutation events to determine whether the category is related to waste sorting deviations, such as a large number of a certain type of waste being incorrectly disposed of.

[0187] The mutation category is used as a key alert category. The frequency of alerts for this category is increased based on the frequency and severity of the deviation, so that users can receive corrective information earlier.

[0188] Based on the types of deviations recorded in behavioral mutation events, the prompt content is adaptively updated. For example, the prompt text is modified to emphasize the correct classification method, or targeted suggestions are added to make the prompt content more consistent with the actual classification deviation.

[0189] During the update of the prompt content, the prompt logic can be optimized based on historical deviation data, so that the updated prompt content can more effectively guide users to correct their classification behavior;

[0190] The system generates a final corrective prompt for waste sorting and continuously pushes it to users during subsequent waste sorting activities, encouraging them to adjust their sorting habits in a timely manner and reduce the frequency of sorting errors.

[0191] In a preferred embodiment of the present invention, the short-cycle volatility indicator and the long-cycle stability indicator are time-aligned, the trend difference value between the two is calculated within the common time period after alignment, and a preliminary correlation result is generated based on the trend difference value, including:

[0192] Time reference points are determined based on short-cycle fluctuation indicators and long-cycle stability indicators, and these time reference points are expanded to form a reference time series that describes the time distribution of the two types of indicators.

[0193] Based on the reference time series, short-cycle volatility indicators and long-cycle stability indicators are time-mapped so that the two types of indicators correspond on the same time axis, forming an aligned indicator sequence.

[0194] Based on the alignment indicator sequence, the direction, magnitude, and speed of change of short-cycle volatility indicators and long-cycle stability indicators within the same time period are obtained, and interval differences are extracted to generate trend difference values.

[0195] Based on the trend difference value, the deviation direction, deviation intensity and deviation duration within the interval are correlated to obtain the primary correlation results used to characterize the relationship of cross-cycle behavior change.

[0196] In this embodiment of the invention, by determining time reference points for short-period fluctuation indicators and long-period stability indicators respectively and performing extended processing, a unified time distribution structure covering both types of indicators can be formed, solving the problem that data at different time scales cannot be directly compared. The formation of the reference time series enables the indicators to be analyzed in a consistent time sequence in subsequent steps.

[0197] By performing time mapping on a reference time series, short-cycle volatility indicators and long-cycle stability indicators can correspond on the same time axis, avoiding alignment deviations caused by differences in time scales. This allows the aligned indicators to more accurately reflect actual behavioral changes. The aligned indicator series obtained through mapping provides a stable data correspondence for subsequent difference extraction.

[0198] By extracting the differences in direction, magnitude, and speed of change within a common time period, the structural differences in behavioral changes between the two types of indicators can be comprehensively reflected, enabling the trend difference value to simultaneously reflect the deviation between short-term changes and long-term trends. The multidimensional difference information contained in the trend difference value can provide more sufficient data for the generation of subsequent triggering factors.

[0199] By analyzing the direction, strength, and persistence of trend differences, preliminary correlation results can be generated to distinguish whether behavioral differences are intrinsically related. These preliminary correlation results provide clearer data direction during the trigger factor construction phase, making the generation of trigger factors more data-driven.

[0200] In a preferred embodiment of the present invention, time reference points are determined based on short-cycle fluctuation indicators and long-cycle stability indicators, respectively, and the time reference points are expanded to form a reference time series describing the time distribution of the two types of indicators, specifically including:

[0201] Extract the start and end times of each short cycle from the time records corresponding to the short cycle fluctuation indicator, and use them as the initial time reference point of the short cycle;

[0202] Extract the start and end times of each long cycle from the time records corresponding to the long cycle stability index, and use them as the initial time reference point of the long cycle.

[0203] Since the time spans of short-term and long-term cycles are different, in order to enable the two types of indicators to be aligned and analyzed under the same time structure, the initial reference point is expanded by extending each reference point forward and backward by a preset time range to form a reference time point that includes the complete time interval.

[0204] The expanded time reference points are arranged in chronological order to ensure that there is no overlap or gap between the reference time points, so as to form a reference time series with complete time coverage.

[0205] The reference time series is used in subsequent time mapping steps to enable comparative analysis of different periodic indicators on a unified time axis.

[0206] In a preferred embodiment of the present invention, short-cycle volatility indicators and long-cycle stability indicators are time-mapped according to a reference time series, so that the two types of indicators correspond on the same time axis, forming an aligned indicator sequence, specifically including:

[0207] Read each time point in the reference time series and use it as the target alignment time for short-cycle volatility indicators and long-cycle stability indicators;

[0208] During the time mapping process, the short-cycle indicator value corresponding to the reference time point is determined by finding the short-cycle fluctuation indicator that is adjacent to or contains the reference time point.

[0209] Similarly, by finding long-term stability indicators that are adjacent to or contain the reference time point, the long-term indicator value corresponding to that time point can be determined.

[0210] The short-cycle indicator values ​​and long-cycle indicator values ​​are arranged in chronological order according to the reference time series so that the two types of indicators at corresponding times can form a one-to-one correspondence.

[0211] This ultimately forms an alignment indicator sequence, enabling short-cycle volatility indicators and long-cycle stability indicators to compare their changing trends over the same time period, laying the foundation for subsequent trend difference extraction.

[0212] In a preferred embodiment of the present invention, based on the alignment index sequence, the direction, magnitude, and rate of change of short-cycle volatility indicators and long-cycle stability indicators within a common time period are obtained, and interval difference extraction is performed to generate trend difference values, specifically including:

[0213] Read the short-cycle volatility index values ​​corresponding to consecutive time points in the alignment index sequence, compare the magnitude relationship between adjacent time points, and determine the direction of change of the short-cycle volatility index within the interval.

[0214] The magnitude of the change is judged by the difference in short-cycle fluctuation indicators, and the interval with excessively large fluctuations is marked as a rapid fluctuation interval.

[0215] The same method as for short cycles is used to analyze the stability index of long cycles, and the trend direction, magnitude and rate of change are judged from the changes at adjacent time points;

[0216] Within a common time period, perform interval difference analysis on the direction, magnitude, and speed of short-cycle and long-cycle indicators, including determining whether the trend directions are consistent, whether the magnitudes of change deviate from each other, and whether there is a gap in the speed of change.

[0217] The above differences are transformed into trend difference values ​​through textual descriptions, so that the trend difference values ​​can indicate whether the behavior of the two time scales deviates within the time interval and the degree of deviation.

[0218] The trend difference value serves as the basis for subsequent correlation judgments and is used to identify the relationship between short-term changes and long-term trends.

[0219] In a preferred embodiment of the present invention, the deviation direction, deviation intensity, and deviation duration within the interval are correlated based on the trend difference value to obtain a preliminary correlation result for characterizing the cross-cycle behavior change relationship, specifically including:

[0220] Analyze the directional characteristics recorded in the trend difference values, and identify the deviation direction by comparing whether the short-cycle change direction is consistent with the long-cycle change direction. If the deviation direction is opposite, it indicates that the behavior change has reverse characteristics.

[0221] Based on the deviation magnitude information in the trend difference value, the deviation intensity is identified and classified into mild, moderate or severe, so that the intensity level can more intuitively represent the difference between the two types of behavior.

[0222] Based on the duration of the trend difference value in multiple consecutive alignment time periods, it is determined whether the duration of the deviation is long enough to reflect whether the cross-cycle offset is persistent.

[0223] By combining three characteristics—direction of deviation, intensity of deviation, and duration of deviation—correlation determination is made on trend differences, forming a preliminary correlation result that can describe the relationship between short-term behavioral changes and long-term trends.

[0224] The initial association results are used in the subsequent trigger factor base value generation step, enabling the behavioral feature expression to be further processed in a hierarchical manner.

[0225] In a preferred embodiment of the present invention, the directionality, cumulativeity, and persistence of the difference within the alignment time period are segmented and analyzed based on the trend difference value in the primary correlation results to generate a short-term triggering factor base value for describing short-term burst behavior and a long-term triggering factor base value for describing long-term offset behavior, including:

[0226] The trend difference values ​​are divided directionally according to the direction of change of the trend difference values, forming directional difference segments to represent the direction of trend change;

[0227] Based on the directional difference segment, the trend difference values ​​within the difference segment are accumulated, and a cumulative difference segment is formed to represent the intensity of the cumulative difference based on the characteristics of the cumulative change.

[0228] Based on the duration of the cumulative difference segment within the alignment time period, a persistence analysis is performed on the difference segment to identify the difference segment whose duration exceeds a preset second persistence threshold, and a short-cycle trigger factor base value is generated based on the difference segment.

[0229] Long-term offset analysis is performed on the difference segments that do not meet the continuous threshold. Trend aggregation is performed based on the offset direction, offset magnitude and offset stage characteristics to generate the base value of long-term trigger factor.

[0230] In this embodiment of the invention, by dividing the trend difference values ​​directionally according to their changing direction, the difference values ​​can be distinguished according to different changing trends, thus clearly expressing the directional nature of short-term and long-term changes. The formation of directional difference segments helps subsequent steps to more accurately identify the category characteristics of the differences.

[0231] By accumulating directional difference segments, the degree of difference over a continuous time period can be integrated into a cumulative difference segment, thus reflecting the intensity of the cumulative change in difference. The cumulative difference segment can be used to identify whether the change exhibits a continuous accumulating trend, thereby providing a basis for subsequently distinguishing between short-term sudden changes and long-term shifts.

[0232] By analyzing the duration of cumulative difference segments within the alignment time period, the portion that consistently exceeds a set threshold can be identified, thereby generating a baseline value for short-cycle triggering factors. This baseline value reflects whether short-cycle behavioral changes are sudden, providing a data foundation for subsequent identification of sudden behaviors.

[0233] Trend aggregation processing is performed on the difference segments that do not reach the persistence threshold. This integrates the offset direction, offset magnitude, and stage change characteristics of these difference segments to form a long-term trigger factor base value. This base value can reflect the offset of behavioral changes in the long-term dimension, providing support for subsequent identification of trend anomalies.

[0234] In a preferred embodiment of the present invention, the trend difference value is directionally divided according to the direction of change of the trend difference value to form a directional difference segment for representing the direction of trend change, specifically including:

[0235] Read the records of trend difference values ​​in consecutive aligned time periods and compare the magnitude changes of trend difference values ​​between adjacent time periods to determine whether the trend difference value is rising, falling, or remaining unchanged;

[0236] Based on the determined direction of change, the trend difference values ​​are divided into multiple directional difference segments in chronological order. For example, trend difference values ​​that are continuously in an upward direction are classified into one directional difference segment, and trend difference values ​​that are continuously in a downward direction are classified into another directional difference segment.

[0237] During the division process, when the trend difference value changes from rising to falling or from falling to rising, it is regarded as a point of change of direction, and the difference segment is divided at this point to ensure that the direction is consistent within each directional difference segment.

[0238] Record the directional difference segments that have been divided so that each difference segment can independently represent the trend difference direction of a certain stage, providing a structured basis for subsequent cumulative analysis.

[0239] In a preferred embodiment of the present invention, based on the directional difference segment, the trend difference values ​​within the difference segment are accumulated, and a cumulative difference segment representing the intensity of the difference accumulation is formed based on the cumulative change characteristics, specifically including:

[0240] In each directional difference segment, the trend difference value within that segment is read one by one, and the difference value is accumulated according to the increase or decrease of the difference value, so that the accumulated result can reflect the overall difference intensity within that directional difference segment.

[0241] When performing cumulative processing, the continuous changes in the trend difference value are judged to identify the growth or decrease trend of the difference value within the segment, and changes in the same direction are accumulated so that the cumulative amount can express the total impact of the difference change within the segment.

[0242] When the cumulative result of the trend difference value within a certain directional difference segment reaches a significant level, the segment is marked as a cumulative difference segment to indicate that the intensity of difference change within the segment is relatively high.

[0243] The cumulative features obtained by accumulating each directional difference segment are recorded and organized into cumulative difference segments for subsequent continuous analysis.

[0244] In a preferred embodiment of the present invention, based on the duration of the cumulative difference segment within the alignment time period, a persistence analysis is performed on the difference segment to identify the difference segment whose duration exceeds a preset second persistence threshold, and a short-cycle trigger factor base value is generated based on the difference segment, specifically including:

[0245] Determine the start and end times for each cumulative difference segment, and calculate the duration of that difference segment;

[0246] The duration is compared with a preset duration threshold. When the duration of the cumulative difference segment exceeds a preset second duration threshold, the difference segment is determined to have sufficient duration to reflect short-cycle sudden behavior.

[0247] For the difference segment that meets the continuous threshold condition, the basic value of the short-cycle trigger factor is generated based on the direction, cumulative amount and change characteristics of the trend difference value within the segment, so that the basic value can reflect the characteristics of short-term sudden behavior.

[0248] The generated short-cycle trigger factor base value is recorded and used for the construction of subsequent hierarchical trigger factors.

[0249] In a preferred embodiment of the present invention, long-term offset analysis is performed on the difference segments that do not meet the persistence threshold. Trend aggregation is then performed based on the offset direction, offset amplitude, and offset stage characteristics to generate a long-term trigger factor base value, specifically including:

[0250] Identify the segments of difference that do not reach the sustained threshold and treat them as offset features that may affect long-term behavior, rather than short-term sudden behaviors;

[0251] Read the trend difference value in the difference segment, and identify the directional characteristics of long-term offset based on the direction of the change in the difference value, such as a continuous overall upward or downward trend.

[0252] Based on the magnitude of the change in trend difference values ​​within the difference segment, the strength of the long-term shift is determined so that the shift strength can reflect the scale of long-term behavioral changes.

[0253] Analyze the changes of this difference segment in different time periods, such as the initial shift, stable shift, and later shift, and identify the structural characteristics of long-term shift through the stage changes;

[0254] The offset direction, offset magnitude, and offset stage characteristics are comprehensively processed to form a trend aggregation result, and based on this, a long-term trigger factor base value is generated to represent the long-term offset behavior characteristics.

[0255] The base value of the long-term triggering factor is used to construct subsequent hierarchical triggering factors to characterize long-term behavioral change trends.

[0256] In a preferred embodiment of the present invention, a preset second persistence threshold is used to determine whether the change in the trend difference value has sufficient persistence within a short period, so as to ensure that the basic value of the short-period triggering factor reflects the real sudden behavior, specifically including:

[0257] Based on the historical changes of trend difference values ​​in multiple short-term analysis windows, the duration of the difference values ​​under normal behavior is statistically analyzed to identify the time difference between short-term fluctuations and persistent fluctuations that may occur within short periods.

[0258] According to the statistical results, sudden behaviors within a short period of time usually have obvious continuity, while occasional differences mostly occur in a single or a small number of windows. By distinguishing between such continuous and discontinuous changes, the accuracy of sudden behavior judgment can be effectively improved.

[0259] Cluster analysis was performed on the time periods in which trend differences persisted to determine the distribution of different duration types, such as short duration, medium duration and long duration, and to determine the shortest possible duration of sudden behavior.

[0260] Based on the actual patterns of waste disposal behavior, such as the length of concentrated periods of residents' disposal time and the typical duration of sudden types of disposal behavior, a time value that can accurately reflect the shortest duration of sudden behavior is selected from the above duration types as the preset second duration threshold.

[0261] When the trend difference exceeds the preset second persistence threshold continuously during the alignment period, the difference can be considered to have the persistence of sudden behavior, and thus can be used to generate the base value of short-cycle triggering factor.

[0262] Embodiments of the present invention also provide a waste disposal optimization management system based on classification data mining, the system comprising:

[0263] The data acquisition module is used to acquire waste disposal behavior data uploaded by the waste disposal collection device, and generate an original behavior sequence to represent continuous disposal behavior based on the data. The waste disposal behavior data includes waste disposal amount, waste disposal time and waste disposal type.

[0264] The sequence construction module is used to divide the original behavior sequence into time periods, generate short-cycle behavior data and long-cycle behavior data, and construct short-cycle classification sequences and long-cycle classification sequences respectively.

[0265] The indicator generation module is used to perform category difference processing between adjacent time periods and category aggregation calculation for each long period based on the short-period classification sequence and the long-period classification sequence, to generate short-period volatility indicators and long-period stability indicators.

[0266] The correlation analysis module is used to perform cross-series correlation analysis based on short-cycle volatility indicators and long-cycle stability indicators, generate correlation analysis results, and construct hierarchical triggering factors including short-cycle triggering factors and long-cycle triggering factors based on them.

[0267] The mutation identification module is used to determine the offset difference based on the hierarchical triggering factor. When the offset difference between the short-cycle triggering factor and the long-cycle triggering factor exceeds the preset associated offset threshold, a behavioral mutation event is generated.

[0268] The strategy migration module is used to perform strategy migration processing based on behavioral mutation events, and generate the migrated waste disposal management strategy, which includes an updated waste bin overflow prediction window, an adjusted waste collection priority, and a reset waste disposal classification correction prompt.

[0269] The waste disposal management module is used to perform waste disposal optimization management operations according to the migrated waste disposal management strategy, and output the execution results to the waste disposal management terminal.

[0270] It should be noted that this system is a system corresponding to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0271] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0272] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0273] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for optimizing management of garbage disposal based on classification data mining, characterized in that, The method includes: Acquire waste disposal behavior data uploaded by waste disposal collection devices, and generate an original behavior sequence to represent continuous disposal behavior based on the data; The original behavior sequence is divided into time periods to generate short-cycle behavior data and long-cycle behavior data, and short-cycle classification sequences and long-cycle classification sequences are constructed respectively. Based on the short-cycle classification sequence and the long-cycle classification sequence, the category difference processing of adjacent time periods and the category aggregation calculation of each long-cycle stage are performed to generate short-cycle fluctuation index and long-cycle stability index. Cross-sequence correlation analysis is performed based on short-cycle volatility indicators and long-cycle stability indicators to generate correlation analysis results, and a hierarchical triggering factor including short-cycle triggering factors and long-cycle triggering factors is constructed based on these results. The offset difference is determined based on the hierarchical triggering factor. When the offset difference between the short-cycle triggering factor and the long-cycle triggering factor exceeds the preset associated offset threshold, a behavioral mutation event is generated. Based on the behavioral mutation event, the strategy migration process is executed to generate the migrated waste disposal management strategy, which includes an updated waste bin overflow prediction window, an adjusted waste collection priority, and a reset waste disposal classification correction prompt. Execute optimized waste disposal management operations based on the relocated waste disposal management strategy, and output the execution results to the waste disposal management terminal; The process of constructing hierarchical trigger factors includes: The short-cycle volatility indicator and the long-cycle stability indicator are time-aligned, and the trend difference value between the two is calculated within the common time period after alignment. The initial correlation result is generated based on the trend difference value. Based on the trend difference values ​​in the primary correlation results, the directionality, cumulativeity, and persistence of the difference in the alignment time period are segmented for analysis, generating basic values ​​of short-term triggering factors to describe short-term burst behavior and basic values ​​of long-term triggering factors to describe long-term offset behavior. Based on the structural differences between the base values ​​of short-cycle trigger factors and long-cycle trigger factors, the two are hierarchically reorganized, with the short-cycle trigger factor placed in the rapid response layer and the long-cycle trigger factor placed in the trend identification layer, forming a layered trigger factor with two layers. 2.The trash drop optimization management method based on classification data mining of claim 1, wherein, Based on the short-cycle and long-cycle classification sequences, class difference processing is performed between adjacent time periods, and class aggregation calculation is performed for each long-cycle stage to generate short-cycle volatility indicators and long-cycle stability indicators, including: The short-period classification sequence is segmented according to a fixed-length short-period analysis window. Within each analysis window, the directional comparison, continuity judgment and amplitude accumulation of the category changes in adjacent time periods are performed to generate short-period difference results within the short-period window. Consistency detection is performed on the category change patterns of multiple consecutive analysis windows based on the short-cycle difference results, and the consistency detection results are transformed into short-cycle fluctuation indicators. Long-period classification sequences are grouped according to long-period stages. Within each long-period group, long-period clustering results are generated based on the frequency of category occurrence, duration, and stability of the dominant category within the stage. Long-period stability indices are then calculated based on the long-period clustering results.

3. The waste disposal optimization management method based on classification data mining according to claim 1, characterized in that, Based on the hierarchical triggering factors, offset difference is determined. When the offset difference between the short-cycle triggering factor and the long-cycle triggering factor exceeds a preset correlation offset threshold, a behavioral mutation event is generated, including: Offset source filtering is performed on the short-cycle trigger factors and long-cycle trigger factors of the layered trigger factors respectively. The instantaneous changes caused by a single anomaly in the short-cycle trigger factors are removed, and the unstable offsets caused by short-term fluctuations in the long-cycle trigger factors are excluded, thus generating filtered short-cycle trigger factors and long-cycle trigger factors. Calculate the continuous change in the offset difference between the filtered short-period triggering factor and the long-period triggering factor, and accumulate the continuous change over multiple consecutive time periods to form an offset difference accumulation sequence. A dual-condition determination is made based on the cumulative amount and duration in the cumulative offset sequence. When the cumulative amount exceeds a preset associated offset threshold and the duration exceeds a preset first duration threshold, a behavioral mutation event is generated. The behavioral mutation event includes a mutation category to indicate the mutation type, a mutation amplitude to indicate the mutation intensity, and a mutation time period to indicate the mutation occurrence range.

4. The waste disposal optimization management method based on classification data mining according to claim 1, characterized in that, Based on the behavioral mutation event, a policy migration process is executed to generate a migrated waste disposal management policy. This includes an updated waste bin overflow prediction window, an adjusted waste collection priority, and reset waste sorting correction prompts, including: The prediction window for overflowing trash cans is dynamically expanded or contracted based on the magnitude and type of the behavioral mutation event, so that the time range of the prediction window changes proportionally with the magnitude of the mutation, and an adjusted prediction window is generated. Based on the waste categories involved in behavioral mutation events and their frequency of occurrence, the collection objects are prioritized. By establishing a collection priority enhancement factor, high mutation categories are given higher collection priority, and an adjusted collection order is generated. Based on the waste sorting deviations corresponding to behavioral mutation events, the deviation category is used as a key prompt category. The frequency of push notifications for this category is increased, and the prompt content is adaptively updated to generate updated waste disposal and sorting correction prompts.

5. The waste disposal optimization management method based on classification data mining according to claim 1, characterized in that, The short-cycle volatility indicator and the long-cycle stability indicator are time-aligned. The trend difference between the two is calculated within their common aligned time period, and a preliminary correlation result is generated based on this trend difference, including: Time reference points are determined based on short-cycle fluctuation indicators and long-cycle stability indicators, and these time reference points are expanded to form a reference time series that describes the time distribution of the two types of indicators. Based on the reference time series, short-cycle volatility indicators and long-cycle stability indicators are time-mapped so that the two types of indicators correspond on the same time axis, forming an aligned indicator sequence. Based on the alignment indicator sequence, the direction, magnitude, and speed of change of short-cycle volatility indicators and long-cycle stability indicators within the same time period are obtained, and interval differences are extracted to generate trend difference values. Based on the trend difference value, the deviation direction, deviation intensity and deviation duration within the interval are correlated to obtain the primary correlation results used to characterize the relationship of cross-cycle behavior change.

6. The waste disposal optimization management method based on classification data mining according to claim 1, characterized in that, Based on the trend difference values ​​in the primary correlation results, a segmented analysis is performed on the directionality, cumulativeity, and persistence of the differences within the alignment time period. This generates basic values ​​for short-term triggering factors to describe short-term burst behavior and basic values ​​for long-term offset behavior, including: The trend difference values ​​are divided directionally according to the direction of change of the trend difference values, forming directional difference segments to represent the direction of trend change; Based on the directional difference segment, the trend difference values ​​within the difference segment are accumulated, and a cumulative difference segment is formed to represent the intensity of the cumulative difference based on the characteristics of the cumulative change. Based on the duration of the cumulative difference segment within the alignment time period, a persistence analysis is performed on the difference segment to identify the difference segment whose duration exceeds a preset second persistence threshold, and a short-cycle trigger factor base value is generated based on the difference segment. Long-term offset analysis is performed on the difference segments that do not meet the continuous threshold. Trend aggregation is performed based on the offset direction, offset magnitude and offset stage characteristics to generate the base value of long-term trigger factor.

7. A waste disposal optimization management system based on classification data mining, characterized in that, The system, used in the method of any one of claims 1 to 6, comprises: The data acquisition module is used to acquire waste disposal behavior data uploaded by the waste disposal collection device, and generate an original behavior sequence to represent continuous disposal behavior based on the data. The waste disposal behavior data includes waste disposal amount, waste disposal time and waste disposal type. The sequence construction module is used to divide the original behavior sequence into time periods, generate short-cycle behavior data and long-cycle behavior data, and construct short-cycle classification sequences and long-cycle classification sequences respectively. The indicator generation module is used to perform category difference processing between adjacent time periods and category aggregation calculation for each long period based on the short-period classification sequence and the long-period classification sequence, to generate short-period volatility indicators and long-period stability indicators. The correlation analysis module is used to perform cross-series correlation analysis based on short-cycle volatility indicators and long-cycle stability indicators, generate correlation analysis results, and construct hierarchical triggering factors including short-cycle triggering factors and long-cycle triggering factors based on them. The mutation identification module is used to determine the offset difference based on the hierarchical triggering factor. When the offset difference between the short-cycle triggering factor and the long-cycle triggering factor exceeds the preset associated offset threshold, a behavioral mutation event is generated. The strategy migration module is used to perform strategy migration processing based on behavioral mutation events, and generate the migrated waste disposal management strategy, which includes an updated waste bin overflow prediction window, an adjusted waste collection priority, and a reset waste disposal classification correction prompt. The waste disposal management module is used to perform waste disposal optimization management operations according to the migrated waste disposal management strategy, and output the execution results to the waste disposal management terminal.

8. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to implement the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the method as described in any one of claims 1 to 6.

Images