Intelligent management system for land development and operation information based on big data

Abstract

The present invention relates to the field of land information data management technology, and in particular to an intelligent management system for land development and operation information based on big data. The system includes a memory and a processor, and the processor executes a computer program stored in the memory to implement the following steps: obtaining development and operation information data of different dimensions in each area of ​​the land to be analyzed; obtaining the correlation value between each two dimensions based on the matching results of the development and operation information data of different dimensions in each two areas and the difference between the development and operation information data of each two dimensions, and then determining the correlation dimension of each dimension; obtaining the corresponding weight and constructing the covariance matrix according to the development and operation information data and the correlation value of each dimension and the associated dimension, reducing the dimension of the covariance matrix, and then compressing and storing all the data. The present invention improves the compression effect of the development and operation information data.

Description

Technical Field

[0001] The present invention relates to the technical field of land information data management, and in particular to an intelligent management system for land development and operation information based on big data. Background Art

[0002] As an important basic information, land resources big data plays an important role in the national economy and social development. A large amount of data will be generated in the process of land resource development and management. Land resource information can be managed by analyzing the large amount of data obtained. When storing the acquired land development and management information, the acquired land data is compressed to improve storage efficiency because it contains a large amount of data. When compressing and storing the acquired land data, the acquired land information data can be clustered by the dynamic connected graph clustering algorithm. Compressing the data in each cluster can greatly improve the compression efficiency.

[0003] When clustering the acquired land information data through the connected graph dynamic split clustering algorithm, since the acquired land information data is multidimensional data, when converting the multidimensional data into nodes, the curse of dimensionality may occur because the data points are high-dimensional data. The clustering distances between all data points are close, resulting in inaccurate subsequent clustering cluster division, affecting the subsequent compression efficiency. When clustering multidimensional land development and management information data, PCA dimensionality reduction is generally performed on the acquired multidimensional data. Clustering the reduced-dimensional data can avoid the above-mentioned curse of dimensionality. However, when the existing PCA algorithm is used to reduce the dimensionality of the acquired land development and management information, only the dimensionality reduction of the acquired data at the data level is considered. The information of the data itself and its meaning may be discarded, resulting in overfitting during data dimensionality reduction, affecting the effect after dimensionality reduction, and making the compression effect of land development and management information data poor. Summary of the invention

[0004] In order to solve the problem of poor compression effect of existing methods when compressing land development and management information data, the purpose of the present invention is to provide an intelligent management system for land development and management information based on big data. The technical solutions adopted are as follows:

[0005] The present invention provides a land development and operation information intelligent management system based on big data, the method comprising the following steps:

[0006] Obtain development and operation information data of different dimensions for each area of ​​the land to be analyzed at each collection time within a preset time period;

[0007] Matching the development and operation information data of different dimensions at all collection moments of every two regions within a preset time period, obtaining a similarity evaluation value corresponding to every two regions based on the matching result, and determining the associated region corresponding to each region based on the similarity evaluation value; obtaining the associated value between every two dimensions based on the difference between the development and operation information data of every two dimensions of each region and its corresponding associated region;

[0008] Determine the associated dimension corresponding to each dimension based on the associated value; obtain the weight of each dimension when reducing the dimension according to the development and operation information data of all the collection moments of each dimension and the corresponding associated dimension and the associated value; obtain the covariance matrix based on the weight; use the PCA dimensionality reduction method to reduce the dimension of the covariance matrix to obtain the characteristic dimension vector after the dimension reduction;

[0009] Based on the similarity of the data of the feature dimension vector after dimensionality reduction, all data are compressed and stored.

[0010] Preferably, matching the development and operation information data of different dimensions at all collection moments of every two regions within a preset time period, and obtaining the similarity evaluation values ​​corresponding to every two regions based on the matching results, includes:

[0011] For the development and operation information data of the ath dimension of any region: in chronological order, record the sequence of the development and operation information data of the ath dimension of the region at all acquisition moments within a preset time period as the data sequence of the ath dimension of the region; perform curve fitting on the data sequence to obtain a fitting curve of the ath dimension of the region, obtain the slope corresponding to each data in the data sequence based on the fitting curve, and use the absolute value of the slope as the first eigenvalue of each data;

[0012] For the i-th region and the i+1-th region: perform dynamic time warping on the data sequences of the same dimension of the i-th region and the i+1-th region respectively to obtain each matching pair, wherein each matching pair contains two data of the same dimension; take the absolute value of the difference between the first eigenvalues ​​of the two data in each matching pair as the first difference of each matching pair; based on the first differences of all matching pairs in all dimensions of the i-th region and the i+1-th region, obtain the similarity evaluation value corresponding to the i-th region and the i+1-th region, and the first differences of all matching pairs in all dimensions are negatively correlated with the similarity evaluation value.

[0013] Preferably, determining the associated region corresponding to each region based on the similarity evaluation value includes:

[0014] For the i-th region:

[0015] A region whose similarity evaluation value with the i-th region is greater than a preset similarity threshold is determined as an associated region corresponding to the i-th region.

[0016] Preferably, obtaining the correlation value between each two dimensions according to the difference between the development and operation information data of each area and its corresponding associated area in each two dimensions includes:

[0017] For the ath and bth dimensions:

[0018] For the i-th region: according to the difference between the slope corresponding to each data in the data sequence of the a-th dimension of the i-th region and the slope corresponding to each data in the data sequence of the b-th dimension, obtain the correlation representation value between the a-th dimension and the b-th dimension of the i-th region;

[0019] According to the difference between the correlation representation values ​​of the ath dimension and the bth dimension of each region and its corresponding associated region, the correlation value between the ath dimension and the bth dimension is obtained.

[0020] Preferably, obtaining the correlation characterization value between the ath dimension and the bth dimension in the i-th region according to the difference between the slope corresponding to each data in the data sequence of the ath dimension in the i-th region and the slope corresponding to each data in the data sequence of the b-th dimension includes:

[0021] Calculate the absolute value of the difference between the slope corresponding to the data at each position in the data sequence of the ath dimension in the i-th region and the slope corresponding to the data at the same position in the data sequence of the b-th dimension in the i-th region, as the second difference between the two data;

[0022] The mean of all second differences corresponding to the a-th dimension and the b-th dimension of the ith region is calculated, and the negative correlation normalization result of the mean is used as the correlation representation value between the a-th dimension and the b-th dimension of the ith region.

[0023] Preferably, the following formula is used to calculate the correlation value between the ath dimension and the bth dimension:

[0024]

[0025] Among them, Q a,b represents the association value between the ath dimension and the bth dimension, u i represents the number of associated regions corresponding to the i-th region, M represents the number of regions in the land to be analyzed, and D i (a,b) represents the correlation representation value between the ath dimension and the bth dimension of the i-th region, It represents the average value of the correlation representation values ​​of the a-th dimension and the b-th dimension of all associated regions corresponding to the i-th region, || represents the absolute value sign, and exp() represents an exponential function with a natural constant as the base.

[0026] Preferably, determining the associated dimension corresponding to each dimension based on the associated value includes:

[0027] For the ath dimension:

[0028] A dimension whose correlation value with the a-th dimension is greater than a preset correlation threshold is determined as the correlation dimension corresponding to the a-th dimension.

[0029] Preferably, the following formula is used to calculate the weight of the a-th dimension during dimensionality reduction:

[0030]

[0031] Among them, ω a represents the weight of the a-th dimension when reducing the dimension, m a Indicates the number of associated dimensions corresponding to the a-th dimension, Q ′ a,r Represents the association value between the a-th dimension and its corresponding r-th associated dimension, represents the mean of all data values ​​of the rth associated dimension corresponding to the ath dimension, f r ′ x represents the xth data value of the rth associated dimension corresponding to the ath dimension, X represents the number of data values ​​in any dimension, and f ax represents the xth data value of the ath dimension, It represents the mean of all data values ​​in the ath dimension, || represents the absolute value sign, norm() represents the normalization function, and exp() represents the exponential function with a natural constant as the base.

[0032] Preferably, obtaining a covariance matrix based on the weights comprises:

[0033] The weights are used as weights when calculating the covariance matrix to construct the covariance matrix.

[0034] Preferably, the similarity of the data based on the feature dimension vector after dimensionality reduction, compressing and storing all data, includes:

[0035] Based on the Euclidean distance between the data in the feature dimension vector after dimensionality reduction, the connected graph dynamic split clustering algorithm is used to cluster the data of the feature dimension vector after dimensionality reduction to obtain each cluster cluster;

[0036] The LZ77 compression algorithm is used to compress the data in each cluster to obtain compressed data, and the compressed data is stored.

[0037] The present invention has at least the following beneficial effects:

[0038] The present invention first evaluates the similarity of development and operation information data of different dimensions in each two areas of the land to be analyzed within a preset time period, screens the associated areas corresponding to each area, and the fluctuation similarity between the development and operation information data of each area and its corresponding associated areas is relatively high. Then, according to the difference between the development and operation information data of each two dimensions of each area and its corresponding associated areas, the association value between each two dimensions is obtained, and then the associated dimension corresponding to each dimension is determined. The development and operation information data and the association value at all acquisition moments of each dimension and the corresponding associated dimension are combined to obtain the weight of each dimension during dimensionality reduction, and then a covariance matrix is ​​constructed, so as to avoid that when the PCA algorithm is used to reduce the dimensionality of the acquired land development and operation information data, only the dimensionality reduction of the acquired data at the data level is considered. The information and meaning of the data itself may be discarded, which will cause overfitting during data dimensionality reduction and affect the effect after dimensionality reduction. The present invention improves the compression effect and compression efficiency of the development and operation information data. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 A flowchart of a method executed by an intelligent management system for land development and operation information based on big data provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following is a detailed description of an intelligent management system for land development and operation information based on big data proposed by the present invention in combination with the accompanying drawings and preferred embodiments.

[0042] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0043] The following is a detailed description of a specific solution of an intelligent management system for land development and operation information based on big data provided by the present invention in conjunction with the accompanying drawings.

[0044] Implementation example of intelligent management system for land development and operation information based on big data:

[0045] The specific scenario targeted by this embodiment is: in the process of storing land development and management information data, due to the large amount of data and the large amount of storage space occupied, the land development and management information data needs to be compressed. When compressing the development and management information data, this embodiment takes into account the regional distribution of land information itself, and the overall information data presents a situation where multiple small areas are composed of a whole. The distribution correlation between different regions is analyzed while taking into account the differences in the impact of multi-dimensional data in different regions. The weights of the matrix elements when obtaining the covariance matrix of the dimensional data during PCA dimensionality reduction are corrected, so that the acquired eigenvectors are more in line with the actual situation, making the dimensionality reduction effect better, thereby improving the compression effect of subsequent land development and management information data.

[0046] This embodiment proposes an intelligent management system for land development and operation information based on big data. Figure 1 The specific steps are as follows:

[0047] Step S1, obtaining development and operation information data of different dimensions for each area in the land to be analyzed at each collection time within a preset time period.

[0048] This embodiment obtains the development and operation information data of the land to be analyzed from the database. The storage method of the development and operation information data in the database is to store it in the form of a data table. When the data is stored, the obtained multidimensional data will be stored in multiple tables. The development and operation information data includes: land area, cadastral number, land ownership, population data, environmental data, land use planning data, etc.; each type of data in the development and operation information data is a dimension of data, that is, land information is a dimension of data, cadastral number is a dimension of data, land ownership is a dimension of data, and so on. In specific applications, the implementer can set the type of development and operation information data according to specific circumstances. It should be noted that since the area of ​​the land to be analyzed may be large, the land to be analyzed may contain multiple regions, so the development and operation information data of different dimensions of each region are collected separately. The development and operation information data of each region may change with the change of time, such as population data and environmental data, which will change with the development and utilization of the land. However, some basic information of the land generally does not change in a short time, such as cadastral number, land area, land ownership, etc. In this embodiment, the collection frequency of development and operation information data is set to once a day, that is, the development and operation information data of different dimensions of each area in the land to be analyzed at each collection moment within the preset time period is collected once a day. The preset time period in this embodiment is a set of all historical moments whose time interval with the current moment is less than or equal to the preset time. The preset duration in this embodiment is one year. In specific applications, the implementer can set the data collection frequency and the duration of the preset time period according to specific circumstances. There is a certain amount of text data in the above-mentioned acquired data, so it is necessary to pre-process the above-mentioned acquired text data and convert it into a feature vector. This embodiment uses the TF-IDF algorithm to process the acquired text data and convert it into a feature vector, that is, to obtain the data value of the development and operation information data of each dimension. It should be noted that the development and operation information data mentioned later are all data values.

[0049] So far, this embodiment has obtained the development and operation information data of different dimensions for each area in the land to be analyzed at each collection time within a preset time period.

[0050] Step S2, matching the development and operation information data of different dimensions at all collection moments of every two regions within a preset time period, obtaining similarity evaluation values ​​corresponding to every two regions based on the matching results, and determining the associated regions corresponding to each region based on the similarity evaluation values; obtaining the associated values ​​between every two dimensions based on the differences between the development and operation information data of every two dimensions of each region and its corresponding associated regions.

[0051] There is a certain similarity between the development and operation data of the same dimension of the land to be analyzed, so the subsequent dynamic splitting clustering of the connected graph can be performed based on the similarity between the development and operation data of the same dimension. When clustering multidimensional data, it is necessary to reduce the dimension of the acquired multidimensional data through the PCA algorithm. The similarity between the feature vectors after dimensionality reduction is used as the distance during clustering to avoid the dimensionality curse caused by the acquired data being high-dimensional data. When performing PCA dimensionality reduction on the acquired land information data, the algorithm only obtains feature vectors through the correlation between the numerical values ​​of the data. The information carried by the dimensional data itself and the impact on other dimensions are not considered, which will cause the feature vectors after dimensionality reduction to be overfitted. Clustering overfitted feature vectors will affect the accuracy of its clustering division and reduce its compression efficiency.

[0052] When performing PCA dimensionality reduction on data, only the numerical values ​​of the acquired dimensional data itself are analyzed, and the information meaning of each dimension of the acquired multidimensional data will be ignored, which will lead to overfitting of the acquired eigenvectors and affect the accuracy of dimensionality reduction. When storing the development and operation information data of each area in the land to be analyzed, the data of a period of time in one area is stored. The land development and operation information of the same area at different times can better understand the land development and operation situation. When storing the development and operation information data, each area needs to be analyzed. When obtaining the significance of each dimensional data in the multidimensional data, the influence between the multidimensional data needs to be considered. By analyzing the consistency of the fluctuation trend between the multidimensional data, the correlation between the two dimensions can be determined to determine the influence between the dimensional data. The land information data of different areas obtained are analyzed. The location distribution of different areas may affect the calculation of the correlation of the multidimensional data in the obtained area. When calculating the correlation of the overall data, the difference in the influence of the multidimensional data in each area is considered, so that the relationship between the dimensions of the acquired multidimensional data is more accurate.

[0053] The more similar the fluctuations of multidimensional data in different regions are, the more regions with similar fluctuations there are. When calculating the correlation of the subsequent overall dimensional data, the greater the weight of each region obtained in the region with similar fluctuation distribution. The subsequent dimensional correlation calculation is performed according to the weight, and the multidimensional data obtained in different regions are matched by the DTW algorithm. By analyzing the matching pairs, the difference in the change of dimensional data caused by the different development times of different regions can be avoided. In this embodiment, the development and operation information data of different dimensions at all collection times of each two regions in the land to be analyzed are matched, and the similarity evaluation value corresponding to each two regions is determined based on the matching results.

[0054] Specifically, for the development and operation information data of the ath dimension of any region: in chronological order, the sequence of the development and operation information data of the ath dimension of the region at all acquisition moments within a preset time period is recorded as the data sequence of the ath dimension of the region; the data sequence is curve-fitted to obtain the fitting curve of the ath dimension of the region, and the slope corresponding to each data in the data sequence is obtained based on the fitting curve, and the absolute value of the slope is used as the first eigenvalue of each data; it should be noted that each data corresponds to an eigenvalue. Curve fitting and the method for obtaining the slope of the data points on the curve are both prior arts and will not be elaborated here.

[0055] For the i-th region and the i+1-th region: perform dynamic time warping on the data sequences of the same dimension of the i-th region and the i+1-th region respectively to obtain each matching pair, wherein each matching pair contains two data of the same dimension; since when the dynamic time warping algorithm is used to process the data in the two sequences, there may be a one-to-many or many-to-one situation, for example: the second data in the first sequence matches the second data in the second sequence and the third data in the second sequence, then the second data in the first sequence and the second data in the second sequence constitute a matching pair, and the second data in the first sequence and the third data in the second sequence constitute a matching pair. The absolute value of the difference between the first eigenvalues ​​of the two data in each matching pair is taken as the first difference of each matching pair; based on the first differences of all matching pairs in all dimensions of the i-th region and the i+1-th region, obtain the similarity evaluation value corresponding to the i-th region and the i+1-th region, and the first differences of all matching pairs in all dimensions are negatively correlated with the similarity evaluation value. A negative correlation relationship indicates that the dependent variable will decrease as the independent variable increases, and the dependent variable will increase as the independent variable decreases. It can be a subtraction relationship, a division relationship, etc., which is determined by actual application. As a specific implementation, a specific calculation formula for the similarity evaluation value is given. The specific calculation formula for the similarity evaluation value corresponding to the i-th region and the i+1-th region is:

[0056]

[0057] Among them, W i,i+1 represents the similarity evaluation value corresponding to the i-th region and the i+1-th region, A represents the number of dimensions of development and operation information data, and J a represents the number of matching pairs of the ath dimension between the i-th region and the i+1-th region, ΔK j represents the first difference of the j-th matching pair in the a-th dimension between the i-th region and the i+1-th region, and exp() represents an exponential function with a natural constant as the base.

[0058] By accumulating the differences in the data of each dimension between the i-th region and the i+1-th region, the similarity of the fluctuations in the development and operation data of the i-th region and the i+1-th region is evaluated. The larger the accumulated result, the more dissimilar the fluctuations in the development and operation data of each dimension between the i-th region and the i+1-th region, that is, the smaller the similarity evaluation value corresponding to the i-th region and the i+1-th region.

[0059] By adopting the above method, the similarity evaluation value corresponding to each two areas of the land to be analyzed can be obtained. The larger the similarity evaluation value, the more similar the fluctuation of the development and operation data of the corresponding two areas; therefore, for the i-th area: the area whose similarity evaluation value with the i-th area is greater than the preset similarity threshold is determined as the associated area corresponding to the i-th area. By adopting this method, the associated area corresponding to each area of ​​the land to be analyzed can be obtained. In this embodiment, the preset similarity threshold is 0.75. In specific applications, the implementer can set it according to the specific situation.

[0060] Next, this embodiment will evaluate the correlation between the development and operation information data of each two dimensions of each region according to the difference in the slopes corresponding to the data in the data sequence of each dimension of each region, and obtain the corresponding correlation characterization value.

[0061] Specifically, for the ath and bth dimensions:

[0062] For the ith region: calculate the absolute value of the difference between the slope corresponding to the data at each position in the data sequence of the ath dimension of the ith region and the slope corresponding to the data at the same position in the data sequence of the bth dimension of the ith region, as the second difference between the two data, that is, multiple second differences are obtained. Calculate the mean of all second differences corresponding to the ath dimension and the bth dimension of the ith region, and use the negative correlation normalization result of the mean as the correlation representation value between the ath dimension and the bth dimension of the ith region. The specific calculation formula for the correlation representation value between the ath dimension and the bth dimension of the ith region is:

[0063]

[0064] Among them, D i (a,b) represents the correlation representation value between the ath dimension and the bth dimension of the ith region, N i represents the number of data in the data sequence of each dimension in the i-th region, k az represents the slope corresponding to the zth data in the data sequence of the ath dimension in the ith region, k bzIt represents the slope corresponding to the zth data in the data sequence of the bth dimension in the ith region, || represents the absolute value sign, and exp() represents the exponential function with a natural constant as the base.

[0065] |k az -k bz | represents the second difference between the zth data in the data sequence of the ath dimension in the ith region and the zth data in the data sequence of the bth dimension; It is used to characterize the mean of all second differences corresponding to the a-th dimension and the b-th dimension of the ith region. When the mean is larger, it means that the correlation between the development and operation data of the a-th dimension and the b-th dimension of the ith region is worse, that is, the correlation representation value between the a-th dimension and the b-th dimension of the ith region is smaller.

[0066] By using the above method, the correlation characterization value between the ath dimension and the bth dimension of each region can be obtained. Next, this embodiment will determine the correlation value between the ath dimension and the bth dimension according to the difference between the correlation characterization values ​​of the ath dimension and the bth dimension of each region and its corresponding associated region. The specific calculation formula for the correlation value between the ath dimension and the bth dimension is:

[0067]

[0068] Among them, Q a,b represents the association value between the ath dimension and the bth dimension, u i represents the number of associated regions corresponding to the i-th region, M represents the number of regions in the land to be analyzed, and D i (a,b) represents the correlation representation value between the ath dimension and the bth dimension of the i-th region, It represents the average value of the correlation representation values ​​of the a-th dimension and the b-th dimension of all associated regions corresponding to the i-th region, || represents the absolute value sign, and exp() represents an exponential function with a natural constant as the base.

[0069] The discrete situation is evaluated by the correlation representation value obtained from the region with consistent fluctuations and its own dimensional correlation. The more associated regions the ith region corresponds to, the better the ith region can represent the overall correlation change in the whole. By analyzing the correlation representation value of the dimension of the associated region corresponding to the ith region and its own difference, the larger the value is, the more special the ith region is, and the smaller its weight should be.

[0070] By adopting the above method, the correlation value between every two dimensions can be obtained.

[0071] Step S3, based on the associated value, determine the associated dimension corresponding to each dimension; obtain the weight of each dimension when reducing the dimension according to the development and operation information data of each dimension and the corresponding associated dimension at all collection times and the associated value; obtain the covariance matrix based on the weight; use the PCA dimensionality reduction method to reduce the dimension of the covariance matrix to obtain the characteristic dimension vector after dimensionality reduction.

[0072] In step S2, this embodiment obtains the correlation value between every two dimensions. The larger the correlation value, the stronger the correlation between the development and operation information data of the corresponding two dimensions. Therefore, for the ath dimension: the dimension whose correlation value with the ath dimension is greater than the preset correlation threshold is determined as the correlation dimension corresponding to the ath dimension. Using this method, the correlation dimension corresponding to each dimension can be obtained. In this embodiment, the preset correlation threshold is 0.68. In specific applications, the implementer can set it according to specific circumstances.

[0073] When clustering high-dimensional data of land management and development, high-dimensional data will cause the curse of dimensionality and affect the clustering results. After the acquired data is reduced in dimension through PCA, cluster analysis is performed through the eigenvector. When performing PCA dimensionality reduction, the difference of each dimension can be analyzed to obtain the weight of the dimensional data, but the correlation of multidimensional data will affect the accuracy of differential situation acquisition. The correlation value of multidimensional data is used to perform weighted processing when calculating the difference between dimensional data, so that the information obtained is more in line with the fluctuation of the dimensional data itself.

[0074] Based on this, this embodiment will obtain the weight of each dimension when reducing the dimension according to the development and operation information data and the associated value at all the collection times of each dimension and the corresponding associated dimension. The specific calculation formula for the weight of the a-th dimension when reducing the dimension is:

[0075]

[0076] Among them, ω a represents the weight of the a-th dimension when reducing the dimension, m a Indicates the number of associated dimensions corresponding to the a-th dimension, Q ′ a,r Represents the association value between the a-th dimension and its corresponding r-th associated dimension, represents the mean of all data values ​​of the rth associated dimension corresponding to the ath dimension, f r ′ x represents the xth data value of the rth associated dimension corresponding to the ath dimension, X represents the number of data values ​​in any dimension, and f ax represents the xth data value of the ath dimension, It represents the mean of all data values ​​in the ath dimension, || represents the absolute value sign, norm() represents the normalization function, and exp() represents the exponential function with a natural constant as the base.

[0077] The weight used to reflect the influence of other dimensions on the two data to be differentiated. The larger the value, the smaller the influence of other dimensions, and the higher the credibility of the difference of the obtained data. ′ a,r As the correlation value between two dimensional data, the greater the difference of the associated data, the greater the influence of the difference of this data, the greater the influence of other dimensions obtained, and the lower its credibility. The larger the value of , the more obvious the difference between the data of the ath dimension and the average value of the data of this dimension, and the greater the data weight of this dimension when calculating.

[0078] By adopting the above method, the weight of each dimension during dimensionality reduction can be obtained, and the weight of each dimension during dimensionality reduction can be used as the weight when calculating the covariance matrix, and the covariance matrix can be constructed based on the development and operation data of all dimensions obtained. The process of constructing the covariance matrix is ​​a prior art and will not be described in detail here. The PCA (Principal Component Analysis) dimensionality reduction method is used to reduce the dimensionality of the covariance matrix to obtain the characteristic dimension vector after dimensionality reduction. The PCA dimensionality reduction method is a prior art and will not be described in detail here.

[0079] Step S4, based on the similarity of the data of the feature dimension vector after dimensionality reduction, compress and store all the data.

[0080] This embodiment has obtained the data of the feature dimension vector after dimensionality reduction. Next, all the data will be compressed based on the similarity of the data of the feature dimension vector after dimensionality reduction.

[0081] Specifically, based on the Euclidean distance between the data in the feature dimension vector after dimensionality reduction, the connected graph dynamic split clustering algorithm is used to cluster the data of the feature dimension vector after dimensionality reduction to obtain multiple cluster clusters; the calculation method of the Euclidean distance between data and the connected graph dynamic split clustering algorithm are both existing technologies and will not be elaborated here.

[0082] The fluctuation of data in the same cluster has certain similarities, and the data in different clusters are quite different. The data with similar distribution is compressed. In this embodiment, the LZ77 compression algorithm is used to compress the data in each cluster to construct a dictionary to obtain the compressed data, and the compressed data is stored. When constructing a dictionary for data with similar distribution, the length of the dictionary will be reduced, thereby improving the compression efficiency and compression effect.

[0083] At this point, the compression and storage of land development and management information data to be analyzed have been completed.

[0084] This embodiment first evaluates the similarity of development and operation information data of different dimensions in each two areas of the land to be analyzed within a preset time period, and screens the associated areas corresponding to each area. The fluctuation similarity between the development and operation information data of each area and its corresponding associated areas is relatively high. Then, based on the difference between the development and operation information data of each two dimensions of each area and its corresponding associated areas, the correlation value between each two dimensions is obtained, and then the associated dimension corresponding to each dimension is determined. The development and operation information data and the correlation value at all acquisition moments of each dimension and the corresponding associated dimension are combined to obtain the weight of each dimension during dimensionality reduction, and then a covariance matrix is ​​constructed, so as to avoid only considering the dimensionality reduction of the acquired data at the data level when the PCA algorithm is used to reduce the dimensionality of the acquired land development and operation information data. The information of the data itself and its meaning may be discarded, which will cause overfitting during data dimensionality reduction and affect the effect after dimensionality reduction. This embodiment improves the compression effect and efficiency of the development and operation information data.

[0085] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A land development and operation information intelligent management system based on big data, comprising a memory and a processor, characterized in that: The processor executes the computer program stored in the memory to implement the following steps: Obtaining development and operation information data of different dimensions for each area of ​​the land to be analyzed at each collection time within a preset time period, wherein each type of data in the development and operation information data is data of one dimension; Matching the development and operation information data of different dimensions at all collection moments of every two regions within a preset time period, obtaining a similarity evaluation value corresponding to every two regions based on the matching result, and determining the associated region corresponding to each region based on the similarity evaluation value; obtaining the associated value between every two dimensions based on the difference between the development and operation information data of every two dimensions of each region and its corresponding associated region; Determine the associated dimension corresponding to each dimension based on the associated value; obtain the weight of each dimension when reducing the dimension according to the development and operation information data of all the collection moments of each dimension and the corresponding associated dimension and the associated value; obtain the covariance matrix based on the weight; use the PCA dimensionality reduction method to reduce the dimension of the covariance matrix to obtain the characteristic dimension vector after the dimension reduction; Based on the similarity of the data of the feature dimension vector after dimensionality reduction, all data are compressed and stored; The matching of the development and operation information data of different dimensions at all collection moments of each two regions within a preset time period, and obtaining the similarity evaluation values ​​corresponding to each two regions based on the matching results, includes: For the development and operation information data of the ath dimension of any region: in chronological order, record the sequence of the development and operation information data of the ath dimension of the region at all acquisition moments within a preset time period as the data sequence of the ath dimension of the region; perform curve fitting on the data sequence to obtain a fitting curve of the ath dimension of the region, obtain the slope corresponding to each data in the data sequence based on the fitting curve, and use the absolute value of the slope as the first eigenvalue of each data; For the i-th region and the i+1-th region: perform dynamic time warping on the data sequences of the same dimension of the i-th region and the i+1-th region respectively to obtain each matching pair, wherein each matching pair contains two data of the same dimension; take the absolute value of the difference between the first eigenvalues ​​of the two data in each matching pair as the first difference of each matching pair; based on the first differences of all matching pairs of all dimensions of the i-th region and the i+1-th region, obtain the similarity evaluation value corresponding to the i-th region and the i+1-th region, and the first differences of all matching pairs of all dimensions are negatively correlated with the similarity evaluation value; The method of obtaining the correlation value between each two dimensions according to the difference between the development and operation information data of each area and its corresponding associated area in each two dimensions includes: For the ath and bth dimensions: For the i-th region: according to the difference between the slope corresponding to each data in the data sequence of the a-th dimension of the i-th region and the slope corresponding to each data in the data sequence of the b-th dimension, obtain the correlation representation value between the a-th dimension and the b-th dimension of the i-th region; According to the difference between the correlation representation values ​​of the ath dimension and the bth dimension of each region and its corresponding associated region, the correlation value between the ath dimension and the bth dimension is obtained; The obtaining, according to the difference between the slope corresponding to each data in the data sequence of the ath dimension in the ith region and the slope corresponding to each data in the data sequence of the bth dimension, a correlation representation value between the ath dimension in the ith region and the bth dimension includes: Calculate the absolute value of the difference between the slope corresponding to the data at each position in the data sequence of the ath dimension in the i-th region and the slope corresponding to the data at the same position in the data sequence of the b-th dimension in the i-th region, as the second difference between the two data; Calculate the mean of all second differences corresponding to the a-th dimension and the b-th dimension of the ith region, and use the negative correlation normalization result of the mean as the correlation representation value between the a-th dimension and the b-th dimension of the ith region; The following formula is used to calculate the correlation value between the ath dimension and the bth dimension: Among them, Q a,b represents the association value between the ath dimension and the bth dimension, u i represents the number of associated regions corresponding to the i-th region, M represents the number of regions in the land to be analyzed, and D i (a,b) represents the correlation representation value between the ath dimension and the bth dimension of the i-th region. represents the average value of the correlation representation values ​​of the a-th dimension and the b-th dimension of all associated regions corresponding to the i-th region, || represents the absolute value sign, and exp() represents an exponential function with a natural constant as the base; The determining, based on the association value, the associated dimension corresponding to each dimension includes: For the ath dimension: A dimension whose correlation value with the a-th dimension is greater than a preset correlation threshold is determined as the correlation dimension corresponding to the a-th dimension.

2. The intelligent management system for land development and operation information based on big data according to claim 1 is characterized in that: The determining the associated region corresponding to each region based on the similarity evaluation value includes: For the ith region: A region whose similarity evaluation value with the i-th region is greater than a preset similarity threshold is determined as an associated region corresponding to the i-th region.

3. The intelligent management system for land development and operation information based on big data according to claim 1 is characterized in that: The following formula is used to calculate the weight of the a-th dimension when reducing the dimension: Among them, ω a represents the weight of the a-th dimension when reducing the dimension, m a Indicates the number of associated dimensions corresponding to the a-th dimension, Q ′ a,r Represents the association value between the a-th dimension and its corresponding r-th associated dimension, represents the mean of all data values ​​of the rth associated dimension corresponding to the ath dimension, f r ′ x represents the xth data value of the rth associated dimension corresponding to the ath dimension, X represents the number of data values ​​in any dimension, and f ax represents the xth data value of the ath dimension, It represents the mean of all data values ​​in the ath dimension, || represents the absolute value sign, norm() represents the normalization function, and exp() represents the exponential function with a natural constant as the base.

4. The intelligent management system for land development and operation information based on big data according to claim 1 is characterized in that: The obtaining of a covariance matrix based on the weights comprises: The weights are used as weights when calculating the covariance matrix to construct the covariance matrix.

5. The intelligent management system for land development and operation information based on big data according to claim 1 is characterized in that: The similarity of the data based on the feature dimension vector after dimensionality reduction, compressing and storing all data, includes: Based on the Euclidean distance between the data in the feature dimension vector after dimensionality reduction, the connected graph dynamic split clustering algorithm is used to cluster the data of the feature dimension vector after dimensionality reduction to obtain each cluster cluster; The LZ77 compression algorithm is used to compress the data in each cluster to obtain compressed data, and the compressed data is stored.

Images