Intelligent Analysis Method and System for Intensive Land Use Based on Multidimensional Indicators

Abstract

This invention relates to the field of land intensive use data analysis technology, specifically to an intelligent analysis method and system for land intensive use based on multi-dimensional indicators. The method includes: collecting multi-source indicator data related to various types of land intensive use and constructing a temporal spatial expression structure; marking abnormal influence relationships appearing in the temporal spatial expression structure as causal segments to be repaired; retrieving similar blocks with similar spatiotemporal sequence sets to the target block and establishing a comparison relationship for the causal segments to be repaired; restoring the causal relationship positions and transmission directions corresponding to the missing multi-source indicator data based on the comparison relationship, forming a block causal chain set; determining the limiting factor and extracting the multi-source indicator data upstream of the limiting factor, the interrupted influence relationship, and the restored transmission path, forming a limiting factor attribution record. This invention enables intelligent identification and causal tracing analysis of land intensive use in cases of data loss and abnormal relationships.

Description

Technical Field

[0001] This invention relates to the field of land intensive use data analysis technology, specifically to a method and system for intelligent analysis of land intensive use based on multi-dimensional indicators. Background Technology

[0002] Against the backdrop of the need for territorial spatial governance and optimal allocation of land resources, intensive land use analysis has become an important foundational technology supporting urban planning, industrial layout, and ecological protection. Currently, relevant analytical systems mainly rely on comprehensive evaluation methods based on statistical indicators, spatial analysis methods based on Geographic Information Systems (GIS), and intelligent prediction methods based on machine learning. However, with the diversification of data sources, the refinement of time scales, and the increasing complexity of land use patterns, existing technologies have gradually revealed a series of technical deficiencies in areas such as multidimensional data fusion, causal relationship modeling, and dynamic evaluation accuracy, making it difficult to meet the requirements of intelligent, dynamic, and interpretable intensive land use analysis in the new stage.

[0003] In the practical application of land intensive use analysis, there are two interrelated problems: First, due to differences in time, space, and statistical standards among data from different sources, the meaning of the same indicator shifts in different years or departments, disrupting the originally clear causal relationship. For example, improvements in transportation no longer show a positive impact on land use efficiency in new data, leading to inconsistent analysis results. Second, this instability of causal relationships is further amplified by the data gap effect. When key indicators are missing in some plots, the system cannot reconstruct a complete causal chain, causing the originally ambiguous relationship due to differences in data standards to be completely broken. This makes it impossible to explain the true causes of inefficient land use or to deduce whether improving a certain condition can increase the degree of intensification. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent analysis method and system for intensive land use based on multidimensional indicators, so as to solve the two interrelated problems mentioned in the background art.

[0005] To achieve the above objectives, the technical solution of the present invention is: a smart analysis method for intensive land use based on multi-dimensional indicators, comprising:

[0006] S1. Collect multi-source indicator data related to the intensive use of various types of land, allocate the multi-source indicator data according to the spatial location of the blocks to obtain the multi-source indicator data set of the target blocks, and arrange them in chronological order to form a spatiotemporal sequence set of multi-source indicator data. Record the order of indicator value changes of all data in the spatiotemporal sequence set within adjacent time slices to construct a temporal spatial expression structure.

[0007] S2. Based on the temporal spatial representation structure, compare the abnormal influence relationship between the land use efficiency data of the target block and at least one multi-source indicator data in continuous time slices, mark the abnormal influence relationship that appears in the temporal spatial representation structure as the causal segment to be repaired, and record the set of causal segments to be repaired.

[0008] Among them, the set of causal segments to be repaired is used to represent the location of causal relationships between multi-source indicator data and their transmission direction;

[0009] S3. In the set of causal segments to be repaired, retrieve similar blocks that have similar spatiotemporal sequence sets to the target block, analyze the influence relationship between the same multi-source indicator data and land use efficiency data in the similar blocks, and establish the control relationship of the causal segments to be repaired; under the condition that the multi-source indicator data is missing and the influence relationship in the similar blocks remains continuous, restore the causal relationship position and its transmission direction corresponding to the missing multi-source indicator data according to the control relationship, and form a block causal chain set;

[0010] S4. Along the block causal chain set, compare the multi-source indicator data of the target block and similar blocks in the current time slice to determine the limiting factors that cause the land use efficiency data of the target block to be abnormal. Extract the multi-source indicator data upstream of the limiting factors, the interrupted influence relationship and the restored transmission path to form the limiting factor attribution record, and finally obtain the limiting factors and the limiting factor attribution record.

[0011] Preferably, in S1, the order of index value changes refers to the symbolic sorting of the increase and decrease trends of each index in the time dimension in the spatiotemporal sequence set of the multi-source index data, with adjacent time slices as the order unit, in order to record the continuous change direction and magnitude relationship of each index in the time series.

[0012] The aforementioned temporal spatial representation structure refers to a multi-layered temporal topology network structure constructed with target blocks as nodes in the spatial dimension and the order of indicator changes in adjacent time slices as edges in the temporal dimension. It is used to express the evolutionary relationship and mutual influence path of multi-source indicator data in time and space.

[0013] The method for constructing the temporal spatial representation structure includes: taking the multi-source index data set of the target block as input, extracting the order of index value changes in adjacent time slices according to the time sequence, establishing connection relationships by using the direction of change between indicators in each time slice as directed edges, forming a multi-layer temporal topology network with the spatial location of the target block as the node index, and generating the temporal spatial representation structure.

[0014] Preferably, in step S2, comparing the abnormal influence relationship between the land use efficiency data of the target block and at least one multi-source indicator data in consecutive time slices is as follows: taking the multi-source indicator data and land use efficiency data corresponding to each time slice in the temporal spatial expression structure as input, calculating the correlation symbol relationship between the change direction of the multi-source indicator data and the change direction of the land use efficiency data in adjacent time slices, and tracking whether the correlation symbol relationship remains consistent in consecutive time slices. When the correlation symbol relationship reverses or deviates from the expected correlation threshold in consecutive time slices, it is determined that there is an abnormal influence relationship between the multi-source indicator data and the land use efficiency data.

[0015] The land use efficiency data refers to the comprehensive results of land use output, resource input and spatial utilization status within a unit area of ​​the target block; the abnormal influence relationship refers to the relationship in which the correlation between the changes of multi-source indicator data and land use efficiency data reverses, fluctuates abnormally and deviates from the statistical stability range under continuous time slices, which is used to indicate that the causal dependence between multi-source indicator data within the target block has changed abnormally in the time dimension.

[0016] Preferably, in S2, the causal segment to be repaired refers to the causal path segment between the time slice interval and its corresponding multi-source index pair that has an abnormal influence relationship in the temporal spatial expression structure, used to record the range of abnormal causal relationship between multi-source index data and land use efficiency data in the time dimension; the set of causal segments to be repaired is a structured record set composed of all causal segments to be repaired.

[0017] Preferably, in step S3, retrieving similar blocks that have a similar spatiotemporal sequence set to the target block is specifically as follows: using the spatiotemporal sequence set of the target block as a benchmark, calculating the similarity index between the multi-source indicator data sequence of each candidate block within the same time slice and the corresponding sequence of the target block; when the similarity index is higher than a preset similarity threshold, the corresponding candidate block is determined as a similar block;

[0018] The candidate block refers to all block units that are spatially located in the same analysis area and have complete multi-source indicator data; the similar block refers to block units that are consistent with or similar to the target block in terms of multi-source indicator data structure, time slice coverage, and spatial function type.

[0019] Preferably, in S3, the correspondence of the causal segment to be repaired refers to the corresponding mapping established between the causal dependency between the same multi-source indicator data and land use efficiency data within the same time slice range in the target block and similar blocks, which is used to represent the structural differences between normal causal chains and abnormal causal chains.

[0020] The normal causal chain refers to a causal path sequence in which the causal dependency between multi-source indicator data and land use efficiency data in similar blocks has a consistent association sign relationship under continuous time slices and no abnormal influence relationship appears; the abnormal causal chain refers to a causal path sequence in which there is an abnormal influence relationship between multi-source indicator data and land use efficiency data in the target block.

[0021] The method for establishing the correlation relationship of the causal segment to be repaired is as follows: taking the influence relationship between the same multi-source indicator data and land use efficiency data in similar blocks as a reference, calculate the association symbol relationship, consistency of change direction and time lag difference of the corresponding causal path in the target block; generate a node matching table between the causal segment to be repaired in the target block and the normal causal chain in similar blocks; mark the differences in causal direction, node position and transmission path in the node matching table to form the correlation relationship of the causal segment to be repaired.

[0022] Preferably, in step S3, the specific method for forming the block causal chain set is as follows: Based on the comparison relationship of the causal segments to be repaired, locate the causal path position where the missing multi-source indicator data of the target block is located; copy the node sequence, direction symbol, and time order of the corresponding causal chain in similar blocks to the causal chain template of the target block; insert the missing node and restore its transmission direction with adjacent nodes according to the time slice index and node position difference item; perform consistency verification on the repaired causal path, and if the causal direction symbol is consistent with that of similar blocks and the time lag difference is lower than the threshold, then the path is confirmed to be valid; summarize all valid repair paths to form the block causal chain set.

[0023] Preferably, in step S4, the limiting factor causing the land use efficiency data anomaly in the target block refers to a multi-source indicator node in the block causal chain set that has a significant negative and unstable impact on the land use efficiency data. The limiting factor is determined as follows: based on all multi-source indicator nodes in the block causal chain set, calculate the causal intensity value of each multi-source indicator node and the land use efficiency data in the current time slice; compare the causal intensity value of each multi-source indicator node with its historical average causal intensity; when the causal intensity sign of the multi-source indicator node is reversed and its change exceeds the stability threshold, mark the multi-source indicator node as a candidate limiting factor; perform a comparison of similar nodes in similar blocks for the candidate limiting factor; when the causal intensity of the multi-source indicator node in the target block deviates from the average level of similar blocks and the direction is inconsistent, determine the multi-source indicator node as a limiting factor.

[0024] Preferably, in S4, the factor attribution record refers to a structured data set that records multi-source indicator data upstream of the limiting factor, interrupted influence relationships, and restored transmission paths, with the limiting factor as the core node.

[0025] The multi-source indicator data upstream of the constraint factor refers to the set of multi-source indicator data nodes that are located before the constraint factor node in the block causal chain set and have a direct or indirect transmission relationship with the constraint factor node; the interrupted influence relationship refers to the set of causal edges in the causal path of the target block that are broken and reversed in direction due to the anomaly of the constraint factor node; the restored transmission path refers to the causal path sequence from the upstream node through the constraint factor node to the land use efficiency result node after the normal causal chain structure in similar blocks has been repaired.

[0026] On the other hand, the present invention provides a land intensive use intelligent analysis system based on multidimensional indicators, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the above-described land intensive use intelligent analysis method based on multidimensional indicators.

[0027] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects:

[0028] 1. In this invention, the causal identification mechanism based on the block spatiotemporal sequence structure can still identify the key causal path affecting land use efficiency even when multi-source indicator data is incomplete or abnormal; by comparing the direction and stability of the causal intensity change between multi-source indicators and land use efficiency within a time slice, it can automatically detect and repair abnormal causal relationships, solving the problem that traditional methods cannot accurately determine the cause of changes in land use efficiency under conditions of missing data or noise.

[0029] 2. In this invention, the key indicators that cause abnormal land use efficiency are automatically identified and traced through the similar block comparison repair and limiting factor attribution analysis mechanism. The system restores the causal transmission direction and node position by comparing the causal chain structure of multi-source indicators between different blocks, and assigns abnormal nodes to their upstream influencing sources, fracture paths and repair chains, thereby realizing intelligent analysis of the whole process from discovering anomalies to locating the causes. Attached Figure Description

[0030] Figure 1 This is a flowchart of an embodiment of the present invention. Detailed Implementation

[0031] Example 1, as Figure 1 As shown, the specific implementation steps of the intelligent analysis method for intensive land use based on multi-dimensional indicators proposed in this invention are as follows:

[0032] S1. Collect multi-source indicator data related to the intensive use of various types of land, allocate the multi-source indicator data according to the spatial location of the blocks to obtain the multi-source indicator data set of the target blocks, and arrange them in chronological order to form a spatiotemporal sequence set of multi-source indicator data. Record the order of indicator value changes of all data in the spatiotemporal sequence set within adjacent time slices to construct a temporal spatial expression structure.

[0033] S2. Based on the temporal spatial representation structure, compare the abnormal influence relationship between the land use efficiency data of the target block and at least one multi-source indicator data in continuous time slices, mark the abnormal influence relationship that appears in the temporal spatial representation structure as the causal segment to be repaired, and record the set of causal segments to be repaired.

[0034] Among them, the set of causal segments to be repaired is used to represent the location of causal relationships between multi-source indicator data and their transmission direction;

[0035] S3. In the set of causal segments to be repaired, retrieve similar blocks that have similar spatiotemporal sequence sets to the target block, analyze the influence relationship between the same multi-source indicator data and land use efficiency data in the similar blocks, and establish the control relationship of the causal segments to be repaired; under the condition that the multi-source indicator data is missing and the influence relationship in the similar blocks remains continuous, restore the causal relationship position and its transmission direction corresponding to the missing multi-source indicator data according to the control relationship, and form a block causal chain set;

[0036] S4. Along the block causal chain set, compare the multi-source indicator data of the target block and similar blocks in the current time slice to determine the limiting factors that cause the land use efficiency data of the target block to be abnormal. Extract the multi-source indicator data upstream of the limiting factors, the interrupted influence relationship and the restored transmission path to form the limiting factor attribution record, and finally obtain the limiting factors and the limiting factor attribution record.

[0037] In this embodiment S1, the order of index value changes refers to the symbolic sorting of the increase and decrease trends of each index in the time dimension in the spatiotemporal sequence set of the multi-source index data, with adjacent time slices as the order unit, in order to record the continuous change direction and magnitude relationship of each index in the time series.

[0038] The aforementioned temporal spatial representation structure refers to a multi-layered temporal topology network structure constructed with target blocks as nodes in the spatial dimension and the order of indicator changes in adjacent time slices as edges in the temporal dimension. It is used to express the evolutionary relationship and mutual influence path of multi-source indicator data in time and space.

[0039] The method for constructing the temporal spatial representation structure includes: taking the multi-source index data set of the target block as input, extracting the order of index value changes in adjacent time slices according to the time sequence, establishing connection relationships by using the direction of change between indicators in each time slice as directed edges, forming a multi-layer temporal topology network with the spatial location of the target block as the node index, and generating the temporal spatial representation structure.

[0040] In this embodiment S1, the multi-source indicator data related to intensive land use refers to a multi-dimensional data set that can characterize land use intensity, utilization efficiency, spatial configuration, and resource input. This data includes land use structure, construction land approval, output per unit area, and tax information data from statistical departments; surface development intensity, building density, surface impermeability, and green coverage data from remote sensing monitoring; road accessibility, public service facility coverage, and infrastructure operation indicators data from geographic information systems; and enterprise distribution, industrial activity types, and investment intensity data from industrial registration systems. All types of data have spatial positioning attributes and time identification attributes to support subsequent spatial mapping and time serialization calculations.

[0041] In this embodiment S1, the multi-source indicator data collection method includes calling corresponding indicator data from different databases through government data interfaces, remote sensing monitoring interfaces, and geographic information system interfaces; the collection process adopts data interface standards, format conversion rules, and coding specifications for unified processing according to different data sources; by matching indicator names, spatial coordinates, and time identifiers, a multi-source indicator data input set with a unified structure definition is formed; the multi-source indicator data is allocated according to the spatial location of blocks, which means that based on geographic coordinates or cadastral codes, the collected various indicator data are projected onto the same spatial reference system, and the block unit to which each indicator data belongs is determined according to the division standards of administrative units, planned plots, or regular grid units; the block spatial location is used to uniquely identify land spatial entities and is a spatial index composed of plot boundary coordinates, area range, and spatial code; the target block refers to the block unit selected as the analysis object, and its corresponding multi-source indicator data set consists of all indicator data allocated to this spatial index.

[0042] In this embodiment S1, the arrangement of multi-source indicator data into a spatiotemporal sequence set by chronological order refers to arranging multi-source indicator data within the same target block in ascending order of time identifiers, using the time identifier field as the sequence index, to form a continuously changing data set in the time dimension. The chronological order is the natural ascending order of the time identifier values, used to ensure the logical consistency of each indicator data on the time axis. The spatiotemporal sequence set is a multi-dimensional mapping structure composed of the target block, time identifier, and indicator values, used to simultaneously express the correspondence between the spatial location of the block and the temporal changes. The time slice refers to the smallest time interval unit used for discretization analysis in the time series, which can be a year, quarter, month, or other fixed period unit. Adjacent time slices refer to two consecutive time slices on the time axis, whose time identifiers have a definite chronological order. The setting of adjacent time slices is used to calculate the differences in changes of various indicators within a continuous time period and to provide a time segmentation basis for forming the order of indicator value changes and the spatiotemporal expression structure of the time series.

[0043] In this embodiment S2, the abnormal influence relationship between the land use efficiency data of the target block and at least one multi-source indicator data in continuous time slices is compared as follows: taking the multi-source indicator data and land use efficiency data corresponding to each time slice in the temporal spatial expression structure as input, the correlation symbol relationship between the change direction of the multi-source indicator data and the change direction of the land use efficiency data in adjacent time slices is calculated, and the correlation symbol relationship is tracked in continuous time slices to see if it remains consistent. When the correlation symbol relationship is reversed or deviates from the expected correlation threshold in continuous time slices, it is determined that there is an abnormal influence relationship between the multi-source indicator data and the land use efficiency data.

[0044] The land use efficiency data refers to the comprehensive results of land use output, resource input and spatial utilization status within a unit area of ​​the target block; the abnormal influence relationship refers to the relationship in which the correlation between the changes of multi-source indicator data and land use efficiency data reverses, fluctuates abnormally and deviates from the statistical stability range under continuous time slices, which is used to indicate that the causal dependence between multi-source indicator data within the target block has changed abnormally in the time dimension.

[0045] In this embodiment S2, the land use efficiency data of the target block refers to the block utilization benefit result calculated comprehensively based on indicators such as output per unit area, tax revenue per unit area, industrial density, and construction land utilization rate. It is used to reflect the overall efficiency between land input and output of the block within a certain period. The land use efficiency data is obtained by normalizing the historical statistical data, industrial activity data, and spatial utilization data of the target block to ensure comparability under different indicator dimensions.

[0046] In this embodiment S2, the continuous time slice refers to a sequence of multiple adjacent time slices arranged in a time-series spatial representation structure in a naturally increasing order according to time identifiers, with a fixed time span and non-overlapping time boundaries. The setting of continuous time slices is used to describe the dynamic change trend of multi-source indicator data and land use efficiency data in the time dimension, ensuring that the comparison process can reflect the continuity and continuity of indicator changes. The abnormal influence relationship includes three types of situations: change correlation reversal, change correlation abnormal fluctuation, and change correlation deviating from the statistical stable interval. Change correlation reversal means that the change direction of multi-source indicator data and land use efficiency data changes from positive correlation to negative correlation or from negative correlation to positive correlation in adjacent time slices. Change correlation abnormal fluctuation means that the sign of the indicator correlation coefficient changes frequently in a short period of time in multiple consecutive time slices, causing the causal dependence relationship to lose stability. Change correlation deviating from the statistical stable interval means that the indicator correlation coefficient exceeds the preset confidence threshold range in the continuous time slice, causing abnormal deviation in the correlation between indicators.

[0047] In this embodiment S2, the correlation sign relationship between the direction of change of multi-source indicator data and the direction of change of land use efficiency data within adjacent time slices refers to the sign correspondence formed by calculating the sign values ​​of the increase and decrease directions of the indicators in adjacent time slices. The correlation sign relationship uses "+1" and "-1" to represent the positive and negative trends of the direction of change, and "0" to represent a stable state of change. When the sign of the direction of change of multi-source indicator data is the same as that of the direction of change of land use efficiency data, it indicates a positive correlation; when the signs are opposite, it indicates a negative correlation. The correlation sign relationship is measured by statistically analyzing the sign consistency ratio within consecutive time slices to measure causal stability. When the correlation sign relationship reverses within consecutive time slices, it means that the multi-source indicator data and land use efficiency data are positively or negatively correlated in the previous time slice, but the direction is opposite in the next time slice. When the correlation sign relationship deviates from the expected correlation threshold, it means that the positive and negative sign consistency ratio within consecutive time slices is lower than the set stability threshold, causing the original correlation direction to become unstable. In this case, the system determines the correspondence between the indicator and the land use efficiency data as an abnormal influence relationship and enters the causal segment marking stage.

[0048] In this embodiment S2, the causal segment to be repaired refers to the causal path segment between the time slice interval and its corresponding multi-source index pair that has an abnormal influence relationship in the temporal spatial expression structure. It is used to record the range of abnormal causal relationship between multi-source index data and land use efficiency data in the time dimension. The set of causal segments to be repaired is a structured record set composed of all causal segments to be repaired.

[0049] In this embodiment S2, the time slice interval with an abnormal influence relationship and the causal path segment between the corresponding multi-source indicator pairs refer to the continuous time range in the directed edge sequence connecting the multi-source indicator nodes and the land use efficiency result nodes in the time-series spatial representation structure where an abnormal influence relationship occurs. This causal path segment is used to describe the causal dependency chain from the multi-source indicator nodes to the land use efficiency result nodes in the time dimension. When the dependency chain experiences sign reversal or causal transmission interruption in a certain time slice interval, it constitutes a causal segment to be repaired. The range of abnormal causal relationship between the multi-source indicator data and the land use efficiency data in the time dimension refers to the entire set of time slices and the set of associated indicator nodes corresponding to the abnormal influence relationship in the time-series spatial representation structure. This range is used to identify the interval where the causal dependency is abnormal on the time axis.

[0050] In this embodiment S3, similar blocks with a similar spatiotemporal sequence set to the target block are retrieved, specifically as follows: based on the spatiotemporal sequence set of the target block, the similarity index between the multi-source indicator data sequence of each candidate block within the same time slice and the corresponding sequence of the target block is calculated; when the similarity index is higher than the preset similarity threshold, the corresponding candidate block is determined as a similar block;

[0051] The candidate block refers to all block units that are spatially located in the same analysis area and have complete multi-source indicator data; the similar block refers to block units that are consistent with or similar to the target block in terms of multi-source indicator data structure, time slice coverage, and spatial function type.

[0052] In this embodiment S3, the similarity index is obtained by comparing the consistency of change direction, deviation of change magnitude, and time delay offset of multi-source index data in the time dimension. The consistency of change direction of multi-source index data in the time dimension refers to the proportion of target blocks and candidate blocks where the signs of change directions of each multi-source index data remain consistent within the same time slice, used to measure the synchronicity of the two blocks in the trend direction. The deviation of change magnitude refers to the average absolute value of the difference in change magnitude of similar indicators between the target block and candidate blocks within the same time slice, used to measure the similarity of the degree of index change. The time delay offset refers to the time slice position required in the time series to achieve maximum correlation between the index change sequences of the two blocks. The similarity index is used to measure the degree of alignment of the indicator change process in time. The similarity index includes three components: the consistency coefficient of change direction, the deviation coefficient of change amplitude, and the time delay correction coefficient. The consistency coefficient of change direction is obtained by statistically analyzing the ratio of the number of identical symbols in adjacent time slices to the total number of time slices. The deviation coefficient of change amplitude is obtained by taking the absolute value of the difference between the indicator change amplitude of the target block and the candidate block in each time slice and calculating the average value. The time delay correction coefficient is obtained by calculating the time offset corresponding to the maximum correlation coefficient between the indicator sequences of the two blocks through a sliding time window and normalizing it. The three components are weighted and summed to form a comprehensive similarity index, which is used to characterize the overall similarity between the candidate block and the target block in terms of temporal features.

[0053] In this embodiment S3, the process of calculating the similarity index between the multi-source index data sequence of each candidate block and the corresponding sequence of the target block within the same time slice includes: using the spatiotemporal sequence set of the target block as a benchmark, aligning the spatiotemporal sequence sets of each candidate block to the same time slice range; calculating the consistency coefficient of change direction, the deviation coefficient of change amplitude, and the time delay correction coefficient for each multi-source index data; weighting and synthesizing the three types of coefficients according to preset weights to obtain the comprehensive similarity index between each candidate block and the target block; selecting several blocks with the highest similarity index among all candidate blocks as similar blocks; the preset similarity threshold is determined based on historical sample statistics and model self-calibration results; in the initial stage, the distribution range of the comprehensive similarity index of known high-similarity block samples in the historical period is calculated to determine the lower limit of the similarity threshold under the confidence level; during system operation, the similarity threshold is dynamically adjusted based on the accuracy evaluation of the actual matching results to keep the determination range of similar blocks within a stable number range; when the comprehensive similarity index of a candidate block is greater than the preset similarity threshold, the candidate block is determined to be a similar block to the target block.

[0054] In this embodiment S3, the correspondence of the causal segment to be repaired refers to the corresponding mapping established between the causal dependency between the same multi-source index data and land use efficiency data within the same time slice range in the target block and similar blocks, which is used to represent the structural differences between normal causal chains and abnormal causal chains.

[0055] The normal causal chain refers to a causal path sequence in which the causal dependency between multi-source indicator data and land use efficiency data in similar blocks has a consistent association sign relationship under continuous time slices and no abnormal influence relationship appears; the abnormal causal chain refers to a causal path sequence in which there is an abnormal influence relationship between multi-source indicator data and land use efficiency data in the target block.

[0056] The method for establishing the correlation relationship of the causal segment to be repaired is as follows: taking the influence relationship between the same multi-source indicator data and land use efficiency data in similar blocks as a reference, calculate the association symbol relationship, consistency of change direction and time lag difference of the corresponding causal path in the target block; generate a node matching table between the causal segment to be repaired in the target block and the normal causal chain in similar blocks; mark the differences in causal direction, node position and transmission path in the node matching table to form the correlation relationship of the causal segment to be repaired.

[0057] In this embodiment S3, the process of calculating the association symbol relationship, change direction consistency, and time lag difference of the corresponding causal path in the target block includes: using the time series between the multi-source indicator data and land use efficiency data in the causal segment to be repaired as input, calculating the symbol of the change direction of the indicator values ​​in adjacent time slices and generating a symbol sequence; comparing the symbol sequences of the same indicator pairs in the target block and similar blocks, and calculating the change direction consistency coefficient by the ratio of the number of symbol consistency times to the total number of time slices; calculating the cross-correlation function using the time series of the same indicator pairs in the target block and similar blocks, and determining the time slice offset corresponding to the maximum correlation value as the time lag difference through a sliding time window; the association symbol relationship is used to represent the symbol matching situation of the causal direction between indicators, and the change... Directional consistency is used to measure the stability of the causal direction between two blocks, and the time lag difference is used to quantify the relative time difference of causal action. The process of generating a node matching table between the causal segment to be repaired in the target block and the normal causal chain in similar blocks includes: using the node sequence of the normal causal chain in similar blocks as a reference, searching for the corresponding multi-source indicator nodes in the causal segment to be repaired in the target block according to the time slice order; for the matched nodes, recording their position index, indicator type, and causal direction symbol in the time slice; for unmatched nodes or nodes with opposite symbol directions, marking them as abnormal nodes and recording the time lag difference between nodes; arranging all matched nodes and abnormal nodes in chronological order to form a node matching table, which is used to express the mapping relationship between the target block and similar blocks in the causal structure.

[0058] In this embodiment S3, identifying the differences in causal direction, node position, and transmission path in the node matching table refers to extracting the difference information of the causal chain structure between two blocks from the node matching table. The causal direction difference item refers to the record where the symbol direction is inconsistent between nodes with the same index, used to identify the situation of causal relationship reversal or abnormal direction. The node position difference item refers to the record where the node index is offset in the time slice sequence, used to characterize the change in the time order of causal action. The transmission path difference item refers to the causal edge record that exists in the similar block but is missing or interrupted in the target block, used to identify the broken part of the causal transmission chain or the missing part of the path. Through comprehensive analysis of the three types of difference items, the structural differences between the causal segment to be repaired in the target block and the normal causal chain in the similar block can be clearly identified.

[0059] In this embodiment S3, the specific method for forming a block causal chain set is as follows: Based on the comparison relationship of the causal segments to be repaired, locate the causal path position where the missing multi-source indicator data of the target block is located; copy the node sequence, direction symbol, and time order of the corresponding causal chain in similar blocks to the causal chain template of the target block; insert the missing node and restore its transmission direction with adjacent nodes according to the time slice index and node position difference item; perform consistency verification on the repaired causal path, and if the causal direction symbol is consistent with that of similar blocks and the time lag difference is lower than the threshold, the path is confirmed to be valid; summarize all valid repair paths to form a block causal chain set.

[0060] In this embodiment S3, the condition of missing multi-source indicator data refers to the fact that in the spatiotemporal sequence set of the target block, some multi-source indicator data failed to be collected or recorded in one or more consecutive time slices, resulting in missing nodes or broken paths in the temporal spatial expression structure of the indicator; the condition of maintaining the continuity of the influence relationship between similar blocks refers to the fact that the causal direction sign between the corresponding multi-source indicator data and land use efficiency data in the similar blocks remains consistent in consecutive time slices, without direction reversal, sign switching or causal edge breakage, and the indicator has a stable node position and a fixed time lag difference in the causal path of the similar blocks; when the above conditions are met simultaneously, it is considered that the missing data of the target block can be repaired through the causal chain structure of the similar blocks.

[0061] In this embodiment S3, the causal relationship position refers to the topological connection position between multi-source indicator nodes and land use efficiency result nodes in the temporal spatial representation structure, used to represent the relative hierarchy of multi-source indicator data in the causal path; the causal relationship position includes three categories: starting node position, intermediate transmission node position, and ending node position, used to describe the structural order of information transmission from upstream indicators to downstream results; the starting node position represents the input end of the causal chain, usually corresponding to external environment or macro-constraint indicators; the intermediate transmission node position represents the transition node of causal transmission, corresponding to the regulatory role of intermediate layer indicators; the ending node position represents the causal output end, corresponding to the response node of land use efficiency data; the transmission direction refers to the information in the causal chain. The transmission direction from upstream to downstream nodes represents the path of multi-source indicator data on land use efficiency data. The transmission direction is recorded in the causal path as directed edges, and its direction is determined by the positive or negative sign of the causal relationship. When the causal direction sign is positive, it indicates that the increase of upstream indicators causes a synchronous increase in downstream results; when the causal direction sign is negative, it indicates that the increase of upstream indicators leads to a decrease in downstream results. The transmission direction not only reflects the positive or negative relationship of causal effects, but also identifies the dominant transmission order of multi-source indicators in the time dimension. When recovering missing data, the system inserts the missing node into the causal path of the target block and restores its transmission direction with adjacent nodes based on the transmission direction sign and time index relationship of the same nodes in similar blocks.

[0062] In this embodiment S4, the limiting factor causing the land use efficiency data anomaly in the target block refers to a multi-source indicator node in the block causal chain set that has a significant negative and unstable impact on the land use efficiency data. The limiting factor is determined as follows: based on all multi-source indicator nodes in the block causal chain set, the causal intensity value of each multi-source indicator node and the land use efficiency data in the current time slice is calculated; the causal intensity value of each multi-source indicator node is compared with its historical average causal intensity; when the sign of the causal intensity of the multi-source indicator node is reversed and its change exceeds the stability threshold, the multi-source indicator node is marked as a candidate limiting factor; the candidate limiting factor is compared with similar nodes in similar blocks; when the causal intensity of the multi-source indicator node in the target block deviates from the average level of similar blocks and the direction is inconsistent, the multi-source indicator node is determined as a limiting factor.

[0063] In this embodiment S4, the multi-source indicator nodes that have a significant negative and unstable impact on land use efficiency data refer to multi-source indicator nodes in the block causal chain set whose causal intensity value is negative in the current time slice and whose change exceeds the stability threshold compared to the historical average causal intensity. Changes in these nodes represent a reversal of the causal direction or a significant weakening of the causal intensity between the indicator and the land use efficiency data, thus limiting the improvement of land use efficiency. The causal intensity value refers to the degree of causal effect of a certain multi-source indicator node on the land use efficiency data in the current time slice; it is obtained by calculating the conditional dependence of the multi-source indicator node on the land use efficiency data over time. The calculation method for the causal intensity value is as follows: based on the block causal chain set of the target block, extract the relationship between each multi-source indicator node and the land use efficiency data. The time series data between rate data are used to calculate the magnitude of the impact of multi-source indicator data on changes in land use efficiency data, while keeping other indicators constant. The causal intensity value is obtained by comprehensively calculating the conditional probability difference, time lag adjustment, and causal direction weight. The sign of the causal intensity value is used to indicate the causal direction, and its absolute value is used to indicate the degree of influence. The historical average causal intensity refers to the average intensity of the causal effect of a certain multi-source indicator node on land use efficiency data in multiple historical time slices of the same target block. The calculation method is as follows: using multiple consecutive time slices as samples, the causal intensity value under each time slice is calculated, and then the arithmetic mean is taken as the historical average causal intensity of the indicator node. This value is used to characterize the long-term stable influence level of the indicator and land use efficiency data, and provides a benchmark reference for judging the current causal intensity anomaly.

[0064] In this embodiment S4, when the sign of the causal intensity of a multi-source indicator node is reversed, it means that the sign of the causal intensity value in the current time slice is opposite to the sign of its historical average causal intensity, indicating that the direction of causal action has changed from positive to negative or from negative to positive. When the change in the causal intensity of a multi-source indicator node exceeds the stability threshold, it means that the absolute value of the difference between the current causal intensity value and the historical average causal intensity exceeds the set stability threshold range, indicating that the interaction strength between the indicator and land use efficiency data has fluctuated significantly. Both of the above situations are marked by the system as unstable influence characteristics to identify potential limiting factor nodes. The stability threshold is a control parameter used to determine whether the change in the causal intensity of a multi-source indicator node belongs to the normal fluctuation range. Its value is automatically calculated by the system based on the causal intensity fluctuation samples in the historical period. Specifically, it includes: calculating the variance of the causal intensity of each multi-source indicator node in the historical time slice, and using the square root of the variance as the standard fluctuation range; multiplying the standard fluctuation range by the confidence coefficient to obtain the stability threshold. During system operation, the stability threshold can be dynamically adjusted according to the number of samples and the time span to ensure that the fluctuation judgment standards between different plots and indicators are comparable.

[0065] In this embodiment S4, the factor attribution record refers to a structured data set that records the multi-source indicator data upstream of the limiting factor, the interrupted influence relationship, and the restored transmission path, with the limiting factor as the core node.

[0066] The multi-source indicator data upstream of the limiting factor refers to the set of multi-source indicator data nodes that are located before the limiting factor node in the block causal chain set and have a direct or indirect transmission relationship with the limiting factor node, used to characterize the upstream driving source affecting the limiting factor; the interrupted influence relationship refers to the set of causal edges in the causal path of the target block that are broken and reversed in direction due to the anomaly of the limiting factor node, used to reflect the location and type of the information transmission chain interruption; the restored transmission path refers to the causal path sequence from the upstream node through the limiting factor node to the land use efficiency result node after the normal causal chain structure in similar blocks has been repaired.

[0067] Example 2: The intelligent analysis system for intensive land use based on multi-dimensional indicators proposed in this invention is applied to the intelligent analysis method for intensive land use based on multi-dimensional indicators proposed in Example 1. It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the intelligent analysis method for intensive land use based on multi-dimensional indicators in Example 1.

[0068] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for intelligent analysis of land intensive use based on multi-dimensional indicators, characterized in that, The method comprises the following steps: S1, collecting multi-source index data related to multi-type land intensive use, distributing the multi-source index data according to the spatial position of the block to obtain a multi-source index data set of the target block, and arranging the multi-source index data set in time sequence to form a time-space sequence set of the multi-source index data, recording the index value change order of all data in the adjacent time slices in the time-space sequence set, and constructing a time-space expression structure; S2, based on the time-space expression structure, comparing the land use efficiency data of the target block with the abnormal influence relationship of at least one multi-source index data in the continuous time slice, marking the abnormal influence relationship appearing in the time-space expression structure as a to-be-repaired causal section, and recording a to-be-repaired causal section set; Wherein, the to-be-repaired causal section set is used to represent the causal relationship position and its transmission direction between the multi-source index data; S3, in the to-be-repaired causal section set, searching for similar blocks with similar time-space sequence sets of the target block, analyzing the influence relationship between the same multi-source index data and the land use efficiency data in the similar blocks, and establishing a control relationship of the to-be-repaired causal section; under the condition that the multi-source index data is missing and the influence relationship in the similar block is continuous, the causal relationship position and its transmission direction corresponding to the missing multi-source index data are recovered according to the control relationship, and a block causal chain set is formed; S4, along the block causal chain set, comparing the multi-source index data of the target block and the similar block under the current time slice, determining the limiting factor causing the abnormality of the land use efficiency data of the target block, and extracting the multi-source index data upstream of the limiting factor, the interrupted influence relationship and the restored transmission path to form a limiting factor attribution record, and finally obtaining the limiting factor and the limiting factor attribution record; The control relationship of the to-be-repaired causal section refers to the corresponding mapping of the causal dependence relationship between the same multi-source index data and the land use efficiency data in the same time slice range in the target block and the similar block, which is used to represent the structural difference between the normal causal chain and the abnormal causal chain; the control relationship of the to-be-repaired causal section is established as follows: taking the influence relationship of the same multi-source index data and the land use efficiency data in the similar block as a reference, calculating the correlation symbol relationship, the change direction consistency and the time lag difference of the corresponding causal path in the target block; generating a node matching table between the to-be-repaired causal section of the target block and the normal causal chain of the similar block; identifying the difference items of the causal direction, the node position and the transmission path in the node matching table to form the control relationship of the to-be-repaired causal section. 2.The multi-dimensional index-based intelligent analysis method for land intensive use according to claim 1, characterized in that: In S1, the index value change order refers to the symbolic ordering of the increase and decrease trend of each index in the time dimension in the time-space sequence set of the multi-source index data, which is used to record the continuous change direction and amplitude relationship of each index in the time sequence; The time sequence space expression structure refers to a multi-layer time sequence topological network structure constructed with a target block as a node in a space dimension and with an index change order of adjacent time slices as an edge in a time dimension, used for expressing the evolution relationship and mutual influence path of multi-source index data in time and space. The construction method of the time sequence space expression structure includes: taking a multi-source index data set of the target block as input, extracting an index value change order in adjacent time slices in time sequence, establishing a connection relationship with the change direction between indexes in each time slice as a directed edge, and forming a multi-layer time sequence topological network with the spatial position of the target block as a node index to generate the time sequence space expression structure. 3.The method of claim 2, wherein the method further comprises: determining the land use intensity of the land use based on the multi-dimensional index. In S2, the abnormal influence relationship between the land use efficiency data of the target block and at least one multi-source index data in consecutive time slices is as follows: taking the multi-source index data and the land use efficiency data corresponding to each time slice in the time sequence space expression structure as input, calculating the correlation sign relationship between the change direction of the multi-source index data and the change direction of the land use efficiency data in adjacent time slices, and tracking whether the correlation sign relationship remains consistent in consecutive time slices; when the correlation sign relationship reverses and deviates from the expected correlation threshold in consecutive time slices, it is determined that there is an abnormal influence relationship between the multi-source index data and the land use efficiency data. The land use efficiency data refers to the comprehensive result data of land use output, resource input and spatial utilization state per unit area in the target block; the abnormal influence relationship refers to the relationship that the change correlation between the multi-source index data and the land use efficiency data reverses, abnormally fluctuates and deviates from the statistical stable interval in consecutive time slices, used to indicate that the causal dependence between the multi-source index data in the target block abnormally changes in the time dimension. 4.The method of claim 3, wherein the method further comprises: determining the land use intensity of the land use based on the multi-dimensional index. In S2, the to-be-repaired causal section refers to the causal path segment between the time slice interval and the corresponding multi-source index pair with an abnormal influence relationship in the time sequence space expression structure, used to record the abnormal causal relationship range between the multi-source index data and the land use efficiency data in the time dimension; the to-be-repaired causal section set is a structured record set composed of all to-be-repaired causal sections. 5.The multi-dimensional index-based intelligent analysis method for land intensive use according to claim 4, characterized in that: In S3, similar blocks with similar time and space sequence sets of the target block are retrieved as follows: taking the time and space sequence set of the target block as a benchmark, calculating the similarity index between the multi-source index data sequence of each candidate block in the same time slice range and the corresponding sequence of the target block; When the similarity index is higher than a preset similarity threshold, the corresponding candidate block is determined as a similar block; The candidate block refers to all block units in the same analysis area in space with complete multi-source index data; the similar block refers to a block unit that is consistent or approximate to the target block in the multi-source index data structure, time slice coverage range and spatial function type. 6.The multi-dimensional index-based intelligent analysis method for land intensification utilization according to claim 5, characterized in that: The normal causal chain in the S3 refers to a causal path sequence in which the causal dependence relationship between the multi-source index data and the land use efficiency data in the similar block is consistent with the associated symbol relationship and no abnormal influence relationship occurs in the continuous time slice; and the abnormal causal chain refers to a causal path sequence in which there is an abnormal influence relationship between the multi-source index data and the land use efficiency data in the target block. 7.The multi-dimensional index-based intelligent analysis method for land intensification utilization according to claim 6, characterized in that: In the S3, the specific method for forming the block causal chain set is as follows: based on the control relationship of the to-be-repaired causal section, the position of the causal path in which the multi-source index data of the target block is missing is located; the node sequence, direction symbol and time sequence of the corresponding causal chain in the similar block are copied into the causal chain template of the target block; According to the time slice index and the node position difference item, the missing node is inserted and the transmission direction between the adjacent nodes is restored; the repaired causal path is subjected to consistency checking, and if the causal direction symbol is consistent with the similar block and the time lag difference is lower than the threshold value, the path is confirmed to be valid; and all the valid repaired paths are collected to form the block causal chain set. 8.The method of claim 7, wherein the method further comprises: In the S4, the limiting factor causing the abnormal land use efficiency data of the target block refers to a multi-source index node that has a significant negative and unstable influence on the land use efficiency data in the block causal chain set; and the determination method of the limiting factor is as follows: based on all the multi-source index nodes in the block causal chain set, the causal strength value of each multi-source index node and the land use efficiency data in the current time slice is calculated; the causal strength value of each multi-source index node is compared with the historical average causal strength, and when the causal strength symbol of the multi-source index node is reversed and the change amplitude exceeds the stability threshold value, the multi-source index node is marked as a candidate limiting factor; The candidate limiting factor is subjected to the same node comparison in the similar block, and when the causal strength of the multi-source index node in the target block deviates from the average level of the similar block and the direction is inconsistent, the multi-source index node is determined to be the limiting factor. 9.The multi-dimensional index-based intelligent analysis method for land intensification utilization according to claim 8, characterized in that: In the S4, the factor attribution record refers to a structured data set recording the multi-source index data upstream of the limiting factor, the interrupted influence relationship and the restored transmission path with the limiting factor as the core node; The multi-source index data upstream of the limiting factor refers to a multi-source index data node set located before the limiting factor node and having a direct or indirect transmission relationship with the limiting factor node in the block causal chain set; the interrupted influence relationship refers to a causal edge set in which the limiting factor node causes the fracture and direction reversal in the causal path of the target block; and the restored transmission path refers to a causal path sequence from the upstream node to the land use efficiency result node through the limiting factor node based on the normal causal chain structure repaired in the similar block. 10.A land intensive use intelligent analysis system based on multi-dimensional indexes, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, and characterized in that: The processor executes a computer program to implement the multi-dimensional index-based intelligent analysis method for land intensive use according to any one of claims 1-9.

Images