Geospatial data grading evaluation method and device based on decision tree and medium

Through the geospatial data grading evaluation method based on decision tree, the identification indicators, evaluation indicators and decision indicator models are constructed, and the decision tree algorithm is used for automated grading, which solves the problems of inaccurate data grading and inefficient management in the existing technology, and realizes the implementation of the problem of inaccurate data grading and low management efficiency. It improves data security and management efficiency, and is suitable for data sharing and transactions in big data environments.

CN119046394AInactive Publication Date: 2024-11-29INSPUR SOFTWARE TECH CO LTD

Patent Information

Application Number
CN202411514832.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-29
Publication Date
2024-11-29
Estimated Expiration
Not applicable · inactive patent

Smart Images

  • Figure CN119046394A_ABST
    Figure CN119046394A_ABST
Patent Text Reader

Abstract

The invention discloses a geographic space data grading evaluation method and device based on a decision tree and a medium, belongs to the technical field of data processing and information security, and aims to solve the technical problems of how to realize effective geographic space data management and effective data grading in use, ensure data security and improve data use efficiency. According to the technical scheme, the method comprises the following steps: constructing a data grading evaluation index model for geographic space data grading; constructing a data grading object feature library according to the data grading evaluation index model; selecting an existing geographic space data feature library data set to train decision tree models, constructing a plurality of decision tree models by adjusting decision tree parameters, and selecting an optimal division condition; and inputting to-be-graded data into the trained decision tree model. The method is used for carrying out effective data security grading evaluation on geospatial data such as vectors, grids and three-dimensional models, and can effectively carry out scientific grading on the geospatial data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of data processing and information security, and in particular to a decision tree-based geographical space data hierarchical evaluation method, equipment and medium. Background Art

[0002] With the rapid development of geographic information science, geospatial data has become an important foundation for decision support and scientific research analysis in various industries. However, geospatial data is of various types and complex sources, and data security and sharing have become key issues that restrict its effective application. The current data classification management ideas are generally based on data characteristics in big data scenarios, and it is believed that data has security attributes of different dimensions, and then construct indicators such as data security risk level and impact degree for data security evaluation. In reality, the industry is mostly based on the definition of business information and data sensitivity level, and constructs data classification and grading based on data sensitivity rules to guide data classification, or defines data classification standards based on expert scoring rules, but these classification strategies often lack reliable measurement indicators, and the classification evaluation relies too much on subjective experience and judgment, and cannot adapt to the changes and development of evaluation scenarios. Traditional geospatial data classification methods still have many subjective judgments, and often lack in-depth mining and utilization of metadata information and data classification management rules, resulting in inaccurate data classification and low management efficiency. Therefore, it is necessary to build a special security classification model for geospatial data and take corresponding hierarchical protection management measures.

[0003] The problems existing in the prior art are as follows: ① The ability to explore and identify geospatial data metadata is weak. Geospatial data metadata is the basis for data classification, involving the structured description of spatial information and data information contained in the data, such as spatial location, attribute data, spatial topological relationships, time series information, etc. This information requires multi-level exploration, but the current recognition and utilization rate of geospatial data metadata is low; ② Information security and surveying and mapping geographic information industry have many regulations and requirements on the hierarchical management of geospatial data. These regulations provide strategies and basis for the hierarchical management objects, grading methods and impact assessment of geospatial data from different perspectives. However, the implementation of these strategies and methods currently relies more on subjective experience, which consumes a lot of manpower and time costs; ③With the rapid development of the surveying and mapping geographic information industry, geospatial data has transformed from small data single modality to big data multimodality. At the same time, the traditional expert evaluation method for the data security level evaluation of geospatial data has played an important role; however, the evaluation result information in the geospatial data security level evaluation process, as an important expert experience, has not been used in the data security evaluation process, and the growth of massive data has also put forward higher requirements for data security evaluation. It is necessary to make full use of the existing expert evaluation experience to assist in improving the work efficiency of data security evaluation.

[0004] Therefore, how to achieve effective geospatial data management and effective data classification during use, ensure data security, improve data utilization efficiency, and at the same time achieve traceability of security assessments is a technical problem that needs to be solved urgently. Summary of the invention

[0005] The technical task of the present invention is to provide a decision tree-based geospatial data grading evaluation method, equipment and medium to solve the problem of how to achieve effective geospatial data management and effective data grading during use, ensure data security, improve data utilization efficiency, and at the same time achieve traceability of security assessments.

[0006] The technical task of the present invention is achieved in the following way: a method for hierarchical evaluation of geographic spatial data based on a decision tree, the method is as follows: Construct a data classification evaluation index model for geospatial data classification: The data classification evaluation index model is constructed based on industry norms and expert knowledge, that is, it serves as a geospatial data classification evaluation knowledge base; wherein, the data classification evaluation index model includes identification indicators, evaluation indicators and decision indicators; Construct a data classification object feature library according to the data classification evaluation index model: Confirm the data source and specify the collection rules through the geospatial data feature library constructed based on the sampling of geospatial data classification objects, and then perform metadata sampling to obtain the data classification identification feature data of the classification objects; Select an existing geospatial data feature library data set to train a decision tree model, and build multiple decision tree models by adjusting decision tree parameters, and then select the optimal partitioning condition; the decision tree parameters include the maximum depth and the minimum number of samples; the decision tree algorithm includes ID3, C4.5 or CART; The data to be classified is input into the trained decision tree model, and the final data level is determined according to the path and evaluation results of the decision tree model.

[0007] As a preference, the identification index is composed of a number of spatial metadata indexes; the identification index includes metadata content of a number of geographic spatial data or data attribute indexes in a normalized data set; the identification index is defined by quantitative or qualitative description, and is divided into two types according to type: data set and data element; Among them, identification indicators come from data fields, data groups, data regions, data accuracy, data scale, data depth, coverage and importance dimensions; identification indicators come from data set data metadata, and are produced and maintained by geospatial data set objects and data production units in accordance with industry standards.

[0008] As a preferred embodiment, the evaluation index is composed of several data security evaluation indicators, including data security evaluation indicators; the evaluation indicators are formulated by experts based on industry norms, standards and industry experience, and the security factors of data objects are evaluated in a quantitative manner; Evaluation indicators are designed to influence objects. The calculation of evaluation indicators is obtained by one or several identification indicators according to calculation rules, or is determined with the assistance of expert experience.

[0009] Preferably, the decision indicator is a data classification target variable, and the evaluation level of the decision indicator defines the data classification standard according to the customized levels of core data, important data and general data.

[0010] As a preferred method, the geospatial data hierarchical object feature library is constructed as follows: Confirm the data source of geospatial data: The data source selects the database and big data platform that stores geospatial data; databases include geospatial databases represented by OracleSDE and PostGIS and document databases represented by MongoDB and minio; big data platforms include big data platforms represented by HBase and Spark; Formulate data source collection rules: including data set metadata collection rules and data set data element collection rules; among them, data set metadata collection rules are to build the mapping relationship and transfer rules between the corresponding data source identification indicators, evaluation indicators and data set fields in the data source according to the data source type, realize the unified identification and escape of geographic spatial data source data description, and convert the metadata and data source information of the data set into the identification indicator label value in the data classification evaluation indicator model; data set data element rules are to identify the sensitive information in the data record by setting data identification rules, and assist in identifying data sensitivity and data record identification, such as sensitive characters, sensitive values, etc.; data identification rules are to dynamically maintain by building a sensitive character information library, and create, modify and delete sensitive characters and data classification levels; Metadata sampling: According to the data source type, select the appropriate data source connection method, establish a connection with the data source, and extract the selected data set according to the established data source sampling rules. The data source sampling rule information includes metadata sampling, data element sampling, and whether to sample the entire data set; among which, data element sampling selects one or more data fields to match the defined collection and identification rules; data source connection methods include ODBC, MDBC and database driver; the connection information of the data source includes matching items of IP address, port number, account number and / or access method; Creation of a hierarchical object feature library: Execute the database data set metadata sampling rules to obtain data grading identification feature data of the hierarchical objects; wherein the feature data includes various identification indicators and evaluation indicators of the data grading evaluation indicator model.

[0011] Preferably, the decision tree model is constructed as follows: Determine the input nodes of the decision tree: The input nodes of the decision tree correspond to the key factors that affect data classification. In the evaluation index model of data classification, only the identification index is used as the input node; the identification index is used as the root node or initial branch node of the decision tree; Construct the branch structure of the decision tree: according to the different values ​​of the input nodes, the decision tree is further branched; specifically: the data field is subdivided into industrial data, telecommunications data, and financial data; the data scale is divided according to the data storage capacity and the group size; each branch represents a specific factor value range or category; Set the middle node of the decision tree: In the middle layer of the decision tree, add nodes based on data impact analysis. The identification indicators and evaluation indicators in the data classification evaluation indicator model are used as middle nodes. The middle nodes are used to evaluate the objects and degree of impact (such as particularly serious harm, serious harm, general harm) that may be affected once the data is leaked, tampered with, damaged, or illegally obtained, illegally used, or illegally shared. Among them, the evaluation indicator can be terminated if the degree of impact involved is the highest level; Define the output nodes of the decision tree: The output nodes correspond to the final level of data classification, such as core data, important data, general data, etc. The final level of data classification will be determined based on the branch path that the data passes through in the decision tree and the evaluation results of the intermediate nodes; Training and adjusting decision trees: Use data sets with known classification results to train decision trees to ensure their accuracy and reliability; during the training process, adjust branch thresholds and pruning parameters to optimize the structure and performance of decision trees; Decision tree partition optimization strategy: Decision tree node partition evaluation methods include information gain-based partitioning strategies (such as ID3 algorithm), gain rate-based partitioning strategies (such as C4.5 algorithm), Gini impurity-based partitioning strategies (such as CART algorithm) and other heuristic partitioning strategies.

[0012] Preferably, a partitioning strategy based on gain rate is used to select the best splitting attribute in the process of building a decision tree model; a data set with known classification results is used as a sample, as follows: When constructing a decision tree for selection, evaluate the division conditions of each node of the decision tree; Input calculation: a numerical attribute value vector composed of specific numerical attribute values ​​of each data contained in the data set used for learning of the decision tree, a label value vector composed of label values ​​of each data, and a group information vector representing the grouping of each data to each node; wherein, the numerical attribute value vector is composed of identification indicators and evaluation indicators in the data set of the data grading evaluation index model, and the numerical attribute value is used to calculate the pros and cons of the division conditions of each node; the label value vector is the decision indicator in the data set of the data grading evaluation index model, that is, the data grading target value that the data grading evaluation index model can predict or classify; the group information vector represents the grouping information to which each sample point in the data belongs at each node; Frequency calculation: using the numerical attribute value vector, label value vector and group information vector, calculate the first frequency of the data belonging to each group, the second frequency of the data for each label value in each group, the third frequency of the data belonging to the divided group obtained by dividing the group according to the division condition representing the comparison between the numerical attribute value and the threshold value, and the fourth frequency of the data for each label value in the divided group; Evaluation calculation: the evaluation value is calculated using the gain rate according to the first frequency, the second frequency, the third frequency and the fourth frequency, and the evaluation value is used to evaluate the division condition; The obtained optimal division conditions are used as the data classification evaluation index model.

[0013] More preferably, the evaluation value is calculated using the gain ratio according to the first frequency, the second frequency, the third frequency, and the fourth frequency as follows: Calculate information entropy: Information entropy is used to measure the degree of confusion of a data set; for a data set , dataset The calculation formula of information entropy is as follows: ; in, The information entropy of the data set D is calculated by summing the logarithmic proportions of each category and then taking the negative value; Represents the information entropy of the data set D, which is used to measure the degree of confusion of the data set; represents the proportion of the i-th category in the sample set; n represents the total number of categories; Calculate information gain (Gain): Information gain indicates the degree to which the information entropy of a data set is reduced after partitioning according to any attribute; and dataset , the information gain is calculated as follows: ; in, Representation attributes For the dataset The information gain of After partitioning, the data set The degree to which the information entropy is reduced; Representation dataset The information entropy of the data set is used to measure the degree of confusion; Representation attributes The set of all possible values; Representation attributes The corresponding value is A subset of samples; and Represents the number of samples in the subset and the entire data set respectively; Representation Subset Information entropy of Calculate split information (SplitInfo): Split information is used to measure the The number of branches generated during the division and the number of samples contained in each branch; the calculation formula is as follows: ; in, Indicates that according to the attributes The split information generated during the partitioning is used to measure the number of branches and the number of samples contained in each branch; Representation attributes The set of all possible values; Representation attributes The corresponding value is The number of sample subsets; Indicates the number of samples in the entire data set; Representation attributes The corresponding value is The proportion of the sample subset to the entire data set; Represents the logarithmic operation of the above ratio; Calculate the gain ratio (GainRatio): The gain ratio is obtained by dividing the information gain by the split information, which is used to balance the influence of information gain and split information; the calculation formula is as follows: ; in, Representation attributes For the dataset The information gain after partitioning measures the degree of reduction in the information entropy of the data set after partitioning; Indicates that according to the attributes The split information generated during the partition is used to measure the number of branches generated by the partition and the number of samples contained in each branch; Indicates the attribute used for partitioning; Represents the entire data set; Calculate the gain rate of the identification index and evaluation index in the data set of the evaluation index model for data grading, compare the gain rate of each division condition, and select the division condition with the largest gain rate as the optimal division condition, so as to improve the purity of data classification without causing the model to be too complicated, thereby simplifying the model.

[0014] An electronic device comprising: a memory and at least one processor; Wherein, the memory stores a computer program; The at least one processor executes the computer program stored in the memory, so that the at least one processor performs the above-mentioned decision tree-based geospatial data hierarchical evaluation method.

[0015] A computer-readable storage medium stores a computer program, which can be executed by a processor to implement the above-mentioned decision tree-based geospatial data hierarchical evaluation method.

[0016] The decision tree-based geospatial data hierarchical evaluation method, device and medium of the present invention have the following advantages: (1) The present invention is used to perform effective data security classification evaluation on geospatial data such as vectors, rasters, and three-dimensional models, and can effectively and scientifically classify geospatial data, providing an important reference for data security management and use. It is applicable to all kinds of organizations and scenarios that need to process geospatial data, and has a wide range of application value; (ii) The present invention constructs a data classification evaluation index model, uses identification indicators, evaluation indicators and decision indicators to comprehensively evaluate the security and importance of data, and realizes scientific classification of geospatial data. It solves the problems of lack of evaluation model for data classification and unclear basis for security evaluation traceability in the current management and use of geospatial data. Through the geospatial data metadata data classification evaluation index model and the decision tree-based evaluation method, the automatic classification evaluation of geospatial data security is realized; (3) The present invention aims to help organizations and enterprises better manage and use geospatial data, ensure data security, improve data usage efficiency, and provide a reliable classification standard for data sharing, exchange and transaction; (IV) The present invention improves data security: by constructing a data classification evaluation index model, the security and sensitivity of geospatial data can be comprehensively evaluated, thereby helping organizations to formulate data security strategies more accurately, prevent sensitive data leakage, and significantly improve data security; (V) The present invention improves data management efficiency: by scientifically classifying geospatial data, organizations can manage their data resources more effectively; different levels of data can adopt different storage, backup and recovery strategies, optimize resource allocation, reduce management costs and improve management efficiency; (VI) The present invention achieves enhanced decision support: it can provide strong support for the decision-making of the organization, and clear data classification helps the organization to quickly identify key information and provide accurate data basis for strategic planning, risk management, etc.; (VII) The present invention promotes data sharing and trading: Under the premise of protecting data security, clear data classification standards help promote data sharing, exchange and trading. Different organizations can conduct more fair and transparent data exchange based on data levels, promoting the rational use of data resources; (VIII) The present invention is flexible and scalable: the decision tree algorithm used is highly flexible and scalable, and can adapt to the data classification requirements of different scales and organizational structures; at the same time, through the version management function, the data classification evaluation index model can be easily customized and adjusted according to actual needs; (IX) The present invention provides a scientific, efficient and flexible method for grading and evaluating geospatial data, which brings significant beneficial effects to organizations in terms of data security, management, decision support, data sharing and trading. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention is further described below in conjunction with the accompanying drawings.

[0018] Attached Figure 1 It is a schematic diagram of the overall process of the hierarchical evaluation method of geospatial data based on decision tree; Attached Figure 2 The figure is a detailed flowchart of the hierarchical evaluation method of geospatial data based on decision trees. DETAILED DESCRIPTION

[0019] The following detailed description is made of the geographic spatial data hierarchical evaluation method, device and medium based on decision tree of the present invention with reference to the drawings and specific embodiments of the specification.

[0020] Embodiment 1: As attached Figure 1 and attached Figure 2 As shown, this embodiment provides a method for hierarchical evaluation of geographic spatial data based on a decision tree, and the method is specifically as follows: S1. Construct a data classification evaluation index model for geospatial data classification: the data classification evaluation index model is constructed based on industry norms and expert knowledge, that is, it serves as a geospatial data classification evaluation knowledge base; wherein, the data classification evaluation index model includes identification indicators, evaluation indicators and decision indicators; S2. Constructing a data classification object feature library according to the data classification evaluation index model: Confirming the data source and specifying the collection rules through the geospatial data feature library constructed based on the sampling of geospatial data classification objects, and then performing metadata sampling to obtain the data classification identification feature data of the classification objects; S3, selecting an existing geospatial data feature library data set to train a decision tree model, and constructing multiple decision tree models by adjusting decision tree parameters, and then selecting the optimal partitioning condition; wherein the decision tree parameters include a maximum depth and a minimum number of samples; and the decision tree algorithm includes ID3, C4.5 or CART; S4. Input the data to be classified into the trained decision tree model, and determine the final data level according to the path and evaluation results of the decision tree model.

[0021] The identification indicators in step S1 of this embodiment are composed of several spatial metadata indicators; the identification indicators include metadata content of several geographic spatial data or data attribute indicators in the normalized data set; the identification indicators are defined by quantitative or qualitative descriptions, and are divided into two types according to the type: data set and data element; as shown in Table 1: Table 1 Identification indicator level table:

[0022] Among them, identification indicators come from data fields, data groups, data areas, data accuracy, data scale, data depth, coverage and importance dimensions, including but not limited to industry fields, business categories, process links, content themes, unit organizations, people involved, regional divisions, plane accuracy, elevation accuracy, resolution accuracy, data storage capacity, data update timeliness, data indicator dimensions, important indicator content, group coverage, regional coverage, time coverage, etc.; identification indicators are derived from data set data metadata, and are produced and maintained by geographic spatial data set objects and data production units in accordance with industry specifications.

[0023] The evaluation index in step S1 of this embodiment is composed of several data security evaluation indicators, including data security evaluation indicators; the evaluation indicators are formulated by experts based on industry norms, standards and industry experience, and the security factors of data objects are evaluated in a quantitative manner; Evaluation indicators are designed to influence objects. The calculation of evaluation indicators is obtained by one or several identification indicators according to calculation rules, or is determined with the assistance of expert experience.

[0024] Evaluation indicators mainly involve impact objects, including but not limited to economic operation, social order, public interests, organizational rights and interests, personal rights and interests, etc. The calculation of evaluation indicators is obtained by one or several identification indicators according to the calculation rules, and expert experience can also be introduced to assist in the determination.

[0025] The decision indicator in step S1 of this embodiment is a data classification target variable, and the evaluation level of the decision indicator defines the data classification standard according to the custom levels of core data, important data and general data. As shown in Table 2: Table 2 Correspondence table of evaluation indicators and decision indicators:

[0026] The construction of the geospatial data hierarchical object feature library in step S2 of this embodiment is as follows: S201. Confirm the data source of geospatial data: the data source selects the database and big data platform for storing geospatial data; the database includes geospatial databases represented by OracleSDE and PostGIS and document databases represented by MongoDB and minio; the big data platform includes big data platforms represented by HBase and Spark; S202. Formulate data source collection rules: including data set metadata collection rules and data set data element collection rules; wherein, data set metadata collection rules are to construct the mapping relationship and transfer rules between the corresponding data source identification indicators, evaluation indicators and data set fields in the data source according to the data source type, realize the unified identification and escape of geographic spatial data source data description, and convert the metadata and data source information of the data set into the identification indicator label value in the data classification evaluation indicator model; data set data element rules are to identify the sensitive information in the data record by setting data identification rules, and assist in identifying data sensitivity and data record identification, such as sensitive characters, sensitive numerical values, etc.; data identification rules are to dynamically maintain by constructing a sensitive character information library, and to create, modify and delete sensitive characters and data classification levels; S203, metadata sampling: according to the data source type, select the appropriate data source connection method, establish a connection with the data source, and extract the selected data set according to the established data source sampling rules. The data source sampling rule information includes metadata sampling, data element sampling, and whether to sample the entire amount; wherein, data element sampling selects one or more data fields to match the defined collection and identification rules; the data source connection method includes using ODBC, MDBC and database driver; the connection information of the data source includes matching items of IP address, port number, account number and / or access method; S204, creating a hierarchical object feature library: executing the database data set metadata sampling rules to obtain data grading identification feature data of the hierarchical objects; wherein the feature data includes various identification indicators and evaluation indicators of the data grading evaluation indicator model.

[0027] The construction of the decision tree model in step S3 of this embodiment is as follows: S301, determine the input nodes of the decision tree: the input nodes of the decision tree correspond to the key factors affecting the data classification, and in the evaluation index model for data classification, only the identification index is used as the input node; the identification index is used as the root node or initial branch node of the decision tree; S302, constructing a branch structure of a decision tree: further branching the decision tree according to different values ​​of input nodes; specifically: the data field is subdivided into industrial data, telecommunications data, and financial data; the data scale is divided according to the data storage volume and the group size; each branch represents a specific factor value range or category; S303, setting the intermediate nodes of the decision tree: in the middle layer of the decision tree, add nodes based on data impact analysis, and the identification indicators and evaluation indicators in the data classification evaluation indicator model are used as intermediate nodes; the intermediate nodes are used to evaluate the objects and the degree of impact (such as particularly serious harm, serious harm, general harm) that may be affected once the data is leaked, tampered with, damaged, or illegally obtained, illegally used, or illegally shared; among them, the evaluation indicator involving the highest level of impact can be terminated; S304, define the output node of the decision tree: the output node corresponds to the final level of data classification, such as core data, important data, general data, etc.; the final level of data classification will be determined according to the branch path and the evaluation results of the intermediate nodes of the data in the decision tree; S305, training and adjusting the decision tree: using a data set with known classification results to train the decision tree to ensure the accuracy and reliability of the decision tree; during the training process, adjusting the branch threshold and pruning parameters to optimize the structure and performance of the decision tree; S306. Decision tree partition optimization strategy: The decision tree node partition evaluation methods include information gain-based partition strategies (such as ID3 algorithm), gain rate-based partition strategies (such as C4.5 algorithm), Gini impurity-based partition strategies (such as CART algorithm) and other heuristic partition strategies.

[0028] The gain rate-based partitioning strategy in step S306 of this embodiment is used to select the best splitting attribute in the process of building a decision tree model; a data set with known classification results is used as a sample, as follows: S30601. When constructing a decision tree for selection, evaluate the conditions for dividing each node of the decision tree; S30602, input calculation: a numerical attribute value vector composed of specific numerical attribute values ​​of each data contained in the data set for learning of the decision tree, a label value vector composed of label values ​​of each data, and a group information vector representing the grouping of each data to each node; wherein the numerical attribute value vector is composed of identification indicators and evaluation indicators in the data set of the data grading evaluation index model, and the numerical attribute value is used to calculate the pros and cons of the division conditions of each node; the label value vector is the decision indicator in the data set of the data grading evaluation index model, that is, the data grading target value that the data grading evaluation index model can predict or classify; the group information vector represents the grouping information to which each sample point in the data belongs at each node; S30603, frequency calculation: using the numerical attribute value vector, the label value vector and the group information vector, calculate the first frequency of the data belonging to each group, the second frequency of the data of each label value in each group, the third frequency of the data belonging to the divided group obtained by dividing the group according to the division condition representing the comparison between the numerical attribute value and the threshold value, and the fourth frequency of the data of each label value in the divided group; S30604, evaluation calculation: using the gain rate to calculate the evaluation value according to the first frequency, the second frequency, the third frequency and the fourth frequency, the evaluation value is used to evaluate the division condition; S30605. Use the obtained optimal division conditions as a data classification evaluation indicator model.

[0029] In step S30604 of this embodiment, the evaluation value is calculated using the gain rate according to the first frequency, the second frequency, the third frequency, and the fourth frequency as follows: S3060401. Calculate information entropy: Information entropy is used to measure the degree of disorder of a data set. , dataset The calculation formula of information entropy is as follows: ; in, The information entropy of the data set D is calculated by summing the logarithmic proportions of each category and then taking the negative value; Represents the information entropy of the data set D, which is used to measure the degree of confusion of the data set; represents the proportion of the i-th category in the sample set; n represents the total number of categories; S3060402. Calculate information gain (Gain): Information gain indicates the degree to which the information entropy of a data set is reduced after partitioning according to any attribute. and dataset , the information gain is calculated as follows: ; in, Representation attributes For the dataset The information gain of After partitioning, the data set The degree to which information entropy is reduced; Representation dataset The information entropy of the data set is used to measure the degree of confusion; Representation attributes The set of all possible values; Representation attributes The corresponding value is A subset of samples; and Represents the number of samples in the subset and the entire data set respectively; Representation Subset Information entropy of S3060403. Calculate split information (SplitInfo): Split information is used to measure the The number of branches generated during the division and the number of samples contained in each branch; the calculation formula is as follows: ; in, Indicates that according to the attributes The split information generated during the partitioning is used to measure the number of branches and the number of samples contained in each branch; Representation attributes The set of all possible values; Representation attributes The corresponding value is The number of sample subsets; Indicates the number of samples in the entire data set; Representation attributes The corresponding value is The proportion of the sample subset to the entire data set; Represents the logarithmic operation of the above ratio; S3060404. Calculate the gain ratio (GainRatio): The gain ratio is obtained by dividing the information gain by the split information, and is used to balance the influence of the information gain and the split information. The calculation formula is as follows: ; in, Representation attributes For the dataset The information gain after partitioning measures the degree of reduction in the information entropy of the data set after partitioning; Indicates that according to the attributes The split information generated during the partition is used to measure the number of branches generated by the partition and the number of samples contained in each branch; Indicates the attribute used for partitioning; Represents the entire data set; S3060405. Calculate the gain rate of the identification index and evaluation index in the data set of the evaluation index model for data grading, compare the gain rate of each division condition, and select the division condition with the largest gain rate as the optimal division condition, so as to improve the purity of data classification without causing the model to be too complicated, thereby simplifying the model.

[0030] Embodiment 2: This embodiment also provides an electronic device, including: a memory and a processor; Wherein, the memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor executes the decision tree-based geospatial data hierarchical evaluation method in any embodiment of the present invention.

[0031] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor may be a microprocessor or any conventional processor, etc.

[0032] The memory can be used to store computer programs and / or modules. The processor realizes various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the terminal, etc. In addition, the memory can also include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SMC), a secure digital (SD) card, a flash memory card, at least one disk storage period, a flash memory device, or other volatile solid-state storage devices.

[0033] Embodiment 3: This embodiment also provides a computer-readable storage medium, in which a plurality of instructions are stored, and the instructions are loaded by a processor, so that the processor executes the decision tree-based geospatial data hierarchical evaluation method in any embodiment of the present invention. Specifically, a system or device equipped with a storage medium can be provided, on which a software program code for implementing the functions of any of the above embodiments is stored, and a computer (or CPU or MPU) of the system or device reads out and executes the program code stored in the storage medium.

[0034] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.

[0035] Examples of storage media for providing program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RYM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, program code may be downloaded from a server computer via a communication network.

[0036] In addition, it should be clear that the functions of any of the above embodiments can be implemented not only by executing the program code read by the computer, but also by enabling an operating system operating on the computer to complete part or all of the actual operations based on instructions from the program code.

[0037] In addition, it can be understood that the program code read from the storage medium is written to a memory provided in an expansion board inserted into the computer or written to a memory provided in an expansion unit connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or the expansion unit is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments.

[0038] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A decision tree-based hierarchical evaluation method for geospatial data, characterized in that: The method is as follows: Construct a data classification evaluation index model for geospatial data classification: the data classification evaluation index model is constructed based on industry norms and expert knowledge, that is, it serves as a geospatial data classification evaluation knowledge base; wherein, the data classification evaluation index model includes identification index, evaluation index and decision index; Construct a data classification object feature library according to the data classification evaluation index model: Confirm the data source and specify the collection rules through the geospatial data feature library constructed based on the sampling of geospatial data classification objects, and then perform metadata sampling to obtain the data classification identification feature data of the classification objects; Select an existing geospatial data feature library data set to train a decision tree model, and build multiple decision tree models by adjusting decision tree parameters, and then select the optimal partitioning condition; where the decision tree parameters include the maximum depth and the minimum number of samples; the decision tree algorithm includes ID3, C4.5 or CART; The data to be classified is input into the trained decision tree model, and the final data level is determined according to the path and evaluation results of the decision tree model.

2. The decision tree-based geospatial data hierarchical evaluation method according to claim 1, characterized in that: Identification indicators include metadata content of some geospatial data or data attribute indicators in normalized data sets; identification indicators are defined through quantitative or qualitative descriptions and are divided into two types: data sets and data elements; Among them, identification indicators come from data fields, data groups, data regions, data accuracy, data scale, data depth, coverage and importance dimensions; identification indicators come from data set metadata, and are produced and maintained by geospatial data set objects and data production units in accordance with industry standards.

3. The decision tree-based geospatial data hierarchical evaluation method according to claim 1, characterized in that: Evaluation indicators include data security evaluation indicators; evaluation indicators are formulated by experts based on industry norms, standards and industry experience, and evaluate the security factors of data objects in a quantitative manner; Evaluation indicators are designed to influence objects. The calculation of evaluation indicators is obtained by one or several identification indicators according to calculation rules, or is determined with the assistance of expert experience.

4. The decision tree-based geospatial data hierarchical evaluation method according to claim 1, characterized in that: Decision indicators are target variables for data classification. The evaluation levels of decision indicators define data classification standards according to the customized levels of core data, important data, and general data.

5. The decision tree-based geospatial data hierarchical evaluation method according to claim 1, characterized in that: The construction of geospatial data hierarchical object feature library is as follows: Confirm the data source of geospatial data: The data source selects the database and big data platform that stores geospatial data; databases include geospatial databases represented by OracleSDE and PostGIS and document databases represented by MongoDB and minio; big data platforms include big data platforms represented by HBase and Spark; Formulate data source collection rules: including data set metadata collection rules and data set data element collection rules; among them, data set metadata collection rules are to build the mapping relationship and transfer rules between the corresponding data source identification indicators, evaluation indicators and data set fields in the data source according to the data source type, realize unified geospatial data source data description recognition and escape, and convert the metadata and data source information of the data set into the identification indicator label value in the data classification evaluation indicator model; data set data element rules are to identify sensitive information in data records by setting data identification rules, and assist in identifying data sensitivity and data record identification; data identification rules are to dynamically maintain by building a sensitive character information library, and to create, modify and delete sensitive characters and data classification levels; Metadata sampling: According to the data source type, select the data source connection method, establish a connection with the data source, and extract the selected data set according to the established data source sampling rules. The data source sampling rule information includes metadata sampling, data element sampling, and whether to sample the entire data set; among which, data element sampling selects one or more data fields to match the defined collection and identification rules; data source connection methods include ODBC, MDBC and database driver; the connection information of the data source includes matching items of IP address, port number, account number and / or access method; Creation of a hierarchical object feature library: Execute the database data set metadata sampling rules to obtain data grading identification feature data of the hierarchical objects; wherein the feature data includes various identification indicators and evaluation indicators of the data grading evaluation indicator model.

6. The method for hierarchical evaluation of geographic spatial data based on a decision tree according to any one of claims 1 to 5, characterized in that: The construction of the decision tree model is as follows: Determine the input nodes of the decision tree: The input nodes of the decision tree correspond to the key factors that affect data classification. In the evaluation index model of data classification, only the identification index is used as the input node; the identification index is used as the root node or initial branch node of the decision tree; Construct the branch structure of the decision tree: according to the different values ​​of the input nodes, the decision tree is further branched; each branch represents a specific factor value range or category; Set the intermediate nodes of the decision tree: In the middle layer of the decision tree, add nodes based on data impact analysis. The identification indicators and evaluation indicators in the data grading evaluation indicator model are used as intermediate nodes. The intermediate nodes are used to evaluate the objects and degree of impact that may be affected if the data is leaked, tampered, damaged, or illegally obtained, used, or shared. Among them, the evaluation indicator can be terminated if the impact level involved is the highest level. Define the output node of the decision tree: The output node corresponds to the final level of data classification, which will be determined based on the branch path that the data passes through in the decision tree and the evaluation results of the intermediate nodes; Training and adjusting decision trees: Use data sets with known classification results to train decision trees to ensure their accuracy and reliability; during the training process, adjust branch thresholds and pruning parameters to optimize the structure and performance of decision trees; Decision tree partition optimization strategy: The decision tree node partition evaluation methods include information gain-based partition strategy, gain rate-based partition strategy, Gini impurity-based partition strategy and heuristic partition strategy.

7. The decision tree-based geospatial data hierarchical evaluation method according to claim 6, characterized in that: The gain rate-based partitioning strategy is used to select the best splitting attribute in the process of building a decision tree model; a data set with known classification results is used as a sample, as follows: When constructing a decision tree for selection, evaluate the division conditions of each node of the decision tree; Input calculation: a numerical attribute value vector composed of specific numerical attribute values ​​of each data contained in the data set used for learning of the decision tree, a label value vector composed of label values ​​of each data, and a group information vector representing the grouping of each data to each node; wherein, the numerical attribute value vector is composed of identification indicators and evaluation indicators in the data set of the data grading evaluation index model, and the numerical attribute value is used to calculate the pros and cons of the division conditions of each node; the label value vector is the decision indicator in the data set of the data grading evaluation index model, that is, the data grading target value that the data grading evaluation index model can predict or classify; the group information vector represents the grouping information to which each sample point in the data belongs at each node; Frequency calculation: using the numerical attribute value vector, label value vector and group information vector, calculate the first frequency of the data belonging to each group, the second frequency of the data for each label value in each group, the third frequency of the data belonging to the divided group obtained by dividing the group according to the division condition representing the comparison between the numerical attribute value and the threshold value, and the fourth frequency of the data for each label value in the divided group; Evaluation calculation: the evaluation value is calculated using the gain rate according to the first frequency, the second frequency, the third frequency and the fourth frequency, and the evaluation value is used to evaluate the division condition; The obtained optimal division conditions are used as the data classification evaluation index model.

8. The decision tree-based geospatial data hierarchical evaluation method according to claim 7, characterized in that: The evaluation value calculated using the gain ratio according to the first frequency, the second frequency, the third frequency, and the fourth frequency is as follows: Calculate information entropy: Information entropy is used to measure the degree of confusion of a data set; for a data set , dataset The calculation formula of information entropy is as follows: ; in, The information entropy of the data set D is calculated by summing the logarithmic proportions of each category and then taking the negative value; Represents the information entropy of the data set D, which is used to measure the degree of confusion of the data set; represents the proportion of the i-th category in the sample set; n represents the total number of categories; Calculate information gain: Information gain indicates the degree to which the information entropy of a data set is reduced after partitioning according to any attribute. and dataset , the information gain is calculated as follows: ; in, Representation attributes For the dataset The information gain of After partitioning, the data set The degree to which information entropy is reduced; Representation dataset The information entropy of the data set is used to measure the degree of confusion; Representation attributes The set of all possible values; Representation attributes The corresponding value is A subset of samples; and Represents the number of samples in the subset and the entire data set respectively; Representation Subset Information entropy of Calculate split information: Split information is used to measure the The number of branches generated during the division and the number of samples contained in each branch; the calculation formula is as follows: ; in, Indicates that according to the attributes The split information generated during the partitioning is used to measure the number of branches and the number of samples contained in each branch; Representation attributes The set of all possible values; Representation attributes The corresponding value is The number of sample subsets; Indicates the number of samples in the entire data set; Representation attributes The corresponding value is The proportion of the sample subset to the entire data set; Represents the logarithmic operation of the above ratio; Calculate the gain rate: The gain rate is obtained by dividing the information gain by the split information, which is used to balance the influence of information gain and split information; the calculation formula is as follows: ; in, Representation attributes For the dataset The information gain after partitioning measures the degree of reduction in the information entropy of the data set after partitioning; Indicates that according to the attributes The split information generated during the partition is used to measure the number of branches generated by the partition and the number of samples contained in each branch; Indicates the attribute used for partitioning; Represents the entire data set; Calculate the gain rate of the identification index and evaluation index in the data set of the evaluation index model, compare the gain rate of each division condition, and select the division condition with the largest gain rate as the optimal division condition.

9. An electronic device, characterized in that: include: memory and at least one processor; Wherein, the memory stores a computer program; The at least one processor executes the computer program stored in the memory, so that the at least one processor performs the decision tree-based geospatial data hierarchical evaluation method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which can be executed by a processor to implement the decision tree-based geospatial data hierarchical evaluation method as described in any one of claims 1 to 8.

Citation Information

Patent Citations

  • Usability evaluation method for open geographic space data

    CN105550303A

  • Emergent event classification and grading method, device and system based on decision trees and Bayesian algorithm

    CN107977670A

  • Data asset classification modeling and grading protection method based on artificial intelligence technology

    CN113590698A

  • Sensitive data classification method based on improved ID3 decision tree

    CN115018007A

  • Decision tree model-based data classification method and related equipment

    CN115099875A

Cited By

  • Surveying and mapping system for IP geographic positioning data

    CN119729763A

  • A mapping system for IP geolocation data

    CN119729763B