A multi-dimensional public data classification method

Abstract

The present invention belongs to the technical field of data classification and grading. The present invention discloses a method for grading and categorizing public data based on multiple dimensions, including: setting multiple classification dimensions and specific categories, formulating levels and rating processes; establishing a mapping relationship between public data metadata and categories, marking direct and indirect dimensions; setting keyword groups for indirect dimensions, and determining classification vectors of various categories based on historical classification data; obtaining target data and its metadata, and determining direct dimension categories according to the mapping relationship; extracting features from the target data to obtain feature vectors, and comparing them with indirect dimension classification vectors to determine categories; integrating the category results of direct and indirect dimensions to form classification data of the target data; determining the initial rating level according to the classification data; determining the final level according to the rating process, and storing the classification data and levels. Classification vectors are established through historical data, and classification is automatically performed according to the calculated similarity, thereby improving the efficiency of classification work.

Description

Technical Field

[0001] The present invention relates to the technical field of data classification and grading, and more specifically, to a multi-dimensional public data grading and classification method. Background Art

[0002] With the development of big data technology, the amount of data collected from various sources has increased dramatically. How to effectively manage and utilize this data has become an important issue. Data classification technology is one of the key technologies to solve this problem. By classifying and classifying data, it can not only improve the management efficiency of data, but also ensure the security and privacy of data, and also facilitate the further analysis and application of data.

[0003] Existing data classification technologies usually use manual classification or simple automatic classification. Manual classification requires a lot of manual intervention, is time-consuming and labor-intensive, and prone to errors. Although simple automatic classification can improve classification efficiency, its classification accuracy is usually low; with the increase in data volume and the refinement of data management requirements, the dimensions that need to be considered when classifying data are also increasing. When faced with multi-dimensional classification needs, the following problems of existing data classification technologies become prominent: As the classification dimensions increase, the classification workload also increases, causing the classification work to become more arduous and time-consuming. Existing automatic classification methods usually cannot accurately capture the characteristics and relationships of data, resulting in low classification accuracy. Existing data classification technologies can usually only handle a few classification dimensions and cannot meet multi-dimensional classification needs. Summary of the invention

[0004] In order to overcome the problem of low classification efficiency in the prior art, the present invention proposes a multi-dimensional public data hierarchical classification method to solve the above problem.

[0005] The present invention provides the following technical solutions:

[0006] A multi-dimensional public data classification method, comprising:

[0007] S1. Set multiple classification dimensions and specific categories under each classification dimension, set several levels from low to high, and establish a rating process;

[0008] S2. Establish a mapping relationship between the metadata of public data and specific categories under the classification dimension, and mark the classification dimension as a direct dimension or an indirect dimension based on the mapping relationship;

[0009] S3, setting a corresponding keyword group for each indirect dimension, obtaining public data that has been classified in the past, and determining the classification vector of each specific category under the indirect dimension according to the set keyword group and the public data that has been classified in the past;

[0010] S4. Obtain the public data to be classified and recorded as target data, obtain metadata of the target data, and obtain specific categories under the direct dimension according to the mapping relationship between the metadata and the category;

[0011] S5. Perform feature extraction on the target data according to the keyword group of the indirect dimension to obtain a feature vector of the target data, and determine the specific category of the target data under the indirect dimension according to the feature vector of the target data and the feature vector of each category under the indirect dimension;

[0012] S6. Use the specific categories of the target data under all direct dimensions and the specific categories of the target data under all indirect dimensions to form the classification data of the target data;

[0013] S7. Determine the initial evaluation level of the target data according to the classification data of the target data;

[0014] S8. Determine the level of the target data according to the initial evaluation level and rating process of the target data; store the classification data and level of the target data.

[0015] Preferably, marking the classification dimension as a direct dimension or an indirect dimension based on the mapping relationship includes: marking the classification dimension of a specific category that has a mapping relationship with the metadata as a direct dimension, and marking the classification dimension of a specific category that has no mapping relationship with the metadata as an indirect dimension.

[0016] Preferably, the step of determining the classification vector of each specific category under the indirect dimension according to the set keyword group and the public data that has been classified in history includes:

[0017] S31, obtaining any specific category under the indirect dimension as the analysis category, and obtaining a keyword group corresponding to the indirect dimension as the analysis phrase;

[0018] S32. extract at least two pieces of public data classified as analysis categories from the public data that have been classified in the past;

[0019] S33, performing feature extraction operations on each piece of extracted public data according to the analysis phrase, and obtaining a feature vector corresponding to each piece of public data;

[0020] S34, obtaining a classification vector of the analysis category according to all feature vectors;

[0021] S35. Repeat steps S31 to S34 to obtain the classification vector of each specific category under the indirect dimension.

[0022] Preferably, the step of performing feature extraction operation includes:

[0023] Obtain keyword groups and public data, and obtain the number n of keywords in the keyword group;

[0024] Create an array with n elements and record it as the characteristic array;

[0025] Each element in the feature array is assigned a value using the following formula:

[0026]

[0027] Where, T i represents the i-th element in the feature array, G i represents the number of times the i-th keyword in the keyword group appears in the public data, G o Indicates the number of times the oth keyword in the keyword group appears in the public data, where o is the index of the keyword;

[0028] Use the assigned feature array as the feature vector to complete the feature extraction operation.

[0029] Preferably, the step of obtaining the classification vector of the analysis category according to all the feature vectors includes:

[0030] Obtain all the feature vectors of the analysis category to form a feature vector set; establish a selection vector with the same number of feature vector elements;

[0031] Synchronously traverse each feature vector and selection vector in the feature vector set in order, and extract the elements of all feature vectors at the traversed position during the traversal process;

[0032] Calculate the median of the extracted elements, and use the calculated median as the element value of the selection vector at the traversal position;

[0033] The selected vector array after traversal is used as the classification vector of the analysis category.

[0034] Preferably, the step of determining the specific category of the target data under the indirect dimension according to the feature vector of the target data and the feature vector of each classification under the indirect dimension comprises:

[0035] Get the feature vector of the target data and all classification vectors under the indirect dimension;

[0036] Calculate the similarity between the feature vector of the target data and each classification vector under the indirect dimension respectively;

[0037] The calculation formula of the similarity is as follows:

[0038]

[0039] In the formula, XS represents similarity, a represents the feature vector of the target data, and b represents the classification vector; · represents the dot product operation, ||a|| represents the modulus length of the feature vector of the target data, ||b|| represents the modulus length of the classification vector, and a j represents the jth element value in the feature vector of the target data, b j represents the jth element value in the classification vector, and m represents the total number of elements in the classification vector;

[0040] Obtaining the maximum value of all calculated similarities, and determining whether the maximum value is greater than or equal to a preset similarity threshold;

[0041] If so, the specific category corresponding to the classification vector corresponding to the maximum value is taken as the specific category of the target data under the indirect dimension;

[0042] If not, the specific category of the target data under the indirect dimension is set to empty.

[0043] Preferably, the several levels from low to high are primary data, secondary data, tertiary data, quaternary data, important data and core data.

[0044] Preferably, the step of determining the initial evaluation level of the target data according to the classification data of the target data is:

[0045] Obtain the classification data of the target data, and obtain all specific categories in the classification data;

[0046] For each specific category, obtain the highest level among all levels of the stored public data of the specific category as the corresponding level of the specific category;

[0047] The highest level among the corresponding levels of all specific categories is obtained as the initial evaluation level of the target data.

[0048] Preferably, the rating process includes:

[0049] Set a corresponding judgment indicator for each level;

[0050] Starting from the judgment indicator corresponding to the highest level, determine whether the data to be rated meets the judgment indicator of the current level;

[0051] If the judgment result is yes, the level of the data to be rated is determined to be the level corresponding to the current judgment indicator; if the judgment result is no, continue to judge the next lower level judgment indicator; until the level of the data to be rated is determined.

[0052] Preferably, determining the level of the target data according to the initial evaluation level and the rating process of the target data includes: obtaining the initial evaluation level, starting the rating process with the judgment index corresponding to the initial evaluation level, and determining the level of the target data.

[0053] The present invention provides a multi-dimensional public data classification method, which has the following beneficial effects:

[0054] By establishing a mapping relationship between metadata and classification dimensions, direct dimensions and indirect dimensions are distinguished. For direct dimensions, classification information can be obtained directly from metadata; for indirect dimensions, by analyzing historical data to establish classification vectors for each category under the indirect dimension, the characteristics of each category can be captured more accurately. The similarity calculation formula is used to determine the specific category of the indirect dimension of the data. This method takes into account the direction and size of the vector and can more accurately determine the category of the data. Improves classification efficiency and classification accuracy. By setting multiple classification dimensions according to actual needs and multiple specific categories under each dimension, the system can customize the classification dimensions according to actual needs. It can meet the different dimensional classification needs of different organizations and industries. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The figure is a flow chart of a multi-dimensional public data classification method of the present invention.

[0056] Figure 2 The present invention is a schematic diagram of a classification vector determination process of a multi-dimensional public data hierarchical classification method. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] Example 1

[0059] See also Figure 1 and Figure 2 This embodiment provides a multi-dimensional public data classification method, including:

[0060] S1. Set multiple classification dimensions and specific categories under each classification dimension, set several levels from low to high, and establish a rating process.

[0061] S2. Establish a mapping relationship between the metadata of public data and specific categories under the classification dimension, and mark the classification dimension as a direct dimension or an indirect dimension based on the mapping relationship.

[0062] The marking of classification dimensions as direct dimensions or indirect dimensions based on the mapping relationship includes: marking classification dimensions of specific categories that have a mapping relationship with metadata as direct dimensions, and marking classification dimensions of specific categories that have no mapping relationship with metadata as indirect dimensions.

[0063] In this embodiment, the classification dimensions can be set according to actual classification needs, such as data generation methods, data application fields and other dimensions, and the specific categories under each dimension can also be set according to actual needs. For example, the specific categories under the data generation method dimension can include manually collected data, information system generated data, sensor device generated data, etc., and the specific categories under the data application field dimension can include urban management, business environment, emergency response, etc. The level and rating process can be set according to actual access control and confidentiality requirements.

[0064] In this embodiment, it should be noted that metadata is data about data, which provides information about the data itself, such as the structure, context, quality, source, etc. of the data; in databases and information systems, metadata is a key component for managing and understanding data. The classification dimension of data generation mode can establish a mapping relationship with the data source in the metadata; for example, if the data source in the metadata is marked as a sensor, it can be mapped to the specific classification of "sensing device generates data"; similarly, if the data source is marked as system generated, it can be mapped to the classification of information system generating data. Furthermore, since the specific classification under the classification dimension of data production mode can be directly determined by metadata, it can be marked as a direct dimension. On the contrary, the specific classification under the classification dimension of data application field cannot be directly determined by metadata, so this classification dimension is marked as an indirect dimension.

[0065] S3. Set a corresponding keyword group for each indirect dimension, obtain the public data that has been classified in the history, and determine the classification vector of each specific category under the indirect dimension according to the set keyword group and the public data that has been classified in the history.

[0066] The step of determining the classification vector of each specific category under the indirect dimension according to the set keyword group and the public data that has been classified in history includes:

[0067] S31, obtaining any specific category under the indirect dimension as the analysis category, and obtaining a keyword group corresponding to the indirect dimension as the analysis phrase;

[0068] S32. extract at least two pieces of public data classified as analysis categories from the public data that have been classified in the past;

[0069] S33, performing feature extraction operations on each piece of extracted public data according to the analysis phrase, and obtaining a feature vector corresponding to each piece of public data;

[0070] S34, obtaining a classification vector of the analysis category according to all feature vectors;

[0071] S35, repeating steps S31 to S34 to obtain the classification vector of each specific category under the indirect dimension. The step of performing the feature extraction operation includes:

[0072] Obtain keyword groups and public data, and obtain the number n of keywords in the keyword group;

[0073] Create an array with n elements and record it as the characteristic array;

[0074] Each element in the feature array is assigned a value using the following formula:

[0075]

[0076] Where, T i represents the i-th element in the feature array, G i represents the number of times the i-th keyword in the keyword group appears in the public data, G o Indicates the number of times the oth keyword in the keyword group appears in the public data, where o is the index of the keyword;

[0077] Use the assigned feature array as the feature vector to complete the feature extraction operation.

[0078] The step of obtaining the classification vector of the analysis category according to all the feature vectors comprises:

[0079] Obtain all the feature vectors of the analysis category to form a feature vector set; establish a selection vector with the same number of feature vector elements;

[0080] Synchronously traverse each feature vector and selection vector in the feature vector set in order, and extract the elements of all feature vectors at the traversed position during the traversal process;

[0081] Calculate the median of the extracted elements, and use the calculated median as the element value of the selection vector at the traversal position;

[0082] The selected vector array after traversal is used as the classification vector of the analysis category.

[0083] In this embodiment, taking the indirect dimension of data application field as an example, a keyword group consisting of keywords including but not limited to greening, urban construction, urban appearance, parks, road maintenance, investment, corporate services, innovation, finance, intellectual property rights, rescue, disaster prevention, public safety, etc. can be set for the indirect dimension.

[0084] The classification vector of each specific classification under the dimension of data application field is determined as follows: Take the specific classification of urban management as an example; first, select urban management as the analysis category. The corresponding analysis phrases are: greening, urban construction, urban appearance, park, road maintenance, investment, corporate services, innovation, finance, intellectual property, rescue, disaster prevention, and public safety. From the historical data, extract two public data classified as urban management: Data 1: The municipal department has stepped up urban greening construction, built many new parks, and improved the urban appearance. Data 2: In order to enhance the image of the city, strengthen road maintenance work and carry out urban construction projects.

[0085] Next, feature extraction is performed on these two data. For data 1, the feature array is [0.33, 0, 0.33, 0.33, 0, 0, 0, 0, 0, 0, 0, 0], which means that greening, city appearance, and park appear once each, and other keywords do not appear. For data 2, the feature array is [0, 0.5, 0, 0, 0.5, 0, 0, 0, 0, 0, 0, 0], indicating that urban construction and road maintenance appear once each, and other keywords do not appear. Based on these two feature vectors, the classification vector of the urban management category can be obtained. The selection vector is [0.16, 0.41, 0, 0.41, 0.25, 0, 0, 0, 0, 0, 0, 0], which is obtained by taking the median of the elements at each position. Repeat the above steps to obtain the classification vectors of the business environment and emergency response categories.

[0086] This method can effectively capture the features of different categories and provide a basis for subsequent text classification tasks. It should be noted that the above steps are schematic. In actual use, classification vectors can be obtained based on more historical data to improve classification accuracy. By analyzing more historical data, more accurate and stable classification vectors can be obtained, thereby improving the performance and reliability of the entire classification system.

[0087] S4. Obtain the public data to be classified and recorded as target data, obtain metadata of the target data, and obtain specific categories under the direct dimension according to the mapping relationship between the metadata and the category.

[0088] S5. Perform feature extraction on the target data according to the keyword group of the indirect dimension to obtain a feature vector of the target data, and determine the specific category of the target data under the indirect dimension according to the feature vector of the target data and the feature vector of each classification under the indirect dimension.

[0089] The step of determining the specific category of the target data under the indirect dimension according to the feature vector of the target data and the feature vector of each classification under the indirect dimension comprises:

[0090] Get the feature vector of the target data and all classification vectors under the indirect dimension;

[0091] Calculate the similarity between the feature vector of the target data and each classification vector under the indirect dimension respectively;

[0092] The calculation formula of the similarity is as follows:

[0093]

[0094] In the formula, XS represents similarity, a represents the feature vector of the target data, and b represents the classification vector; · represents the dot product operation, ||a|| represents the modulus length of the feature vector of the target data, ||b|| represents the modulus length of the classification vector, and a j represents the jth element value in the feature vector of the target data, b j represents the jth element value in the classification vector, and m represents the total number of elements in the classification vector;

[0095] Obtaining the maximum value of all calculated similarities, and determining whether the maximum value is greater than or equal to a preset similarity threshold;

[0096] If so, the specific category corresponding to the classification vector corresponding to the maximum value is taken as the specific category of the target data under the indirect dimension;

[0097] If not, the specific category of the target data under the indirect dimension is set to empty.

[0098] S6. Use the specific categories of the target data under all direct dimensions and the specific categories of the target data under all indirect dimensions to form the classification data of the target data.

[0099] In this embodiment, the steps for classifying the public data to be classified are as follows: first, obtain a piece of public data to be classified and record it as target data, obtain the metadata of the target data, and if the data source in the metadata is system generated; according to the preset category mapping relationship, determine that the specific category of the data under the direct dimension data generation method is information system generated data. Next, use the previously defined keyword group to extract features of the target data to obtain a feature vector; use the similarity calculation formula to calculate the similarity between each classification vector under the indirect dimension; the similarity calculation formula measures the similarity between two vectors by two aspects: vector direction and vector value; the higher the similarity value, the more similar the two vectors are.

[0100] Taking the data application field dimension as an example, using the given similarity calculation formula, the following similarity data are calculated: the similarity with urban management is 0.3162, the similarity with the business environment is 0.5774, and the similarity with emergency response is 0. If the similarity threshold is set to 0.5, then the maximum similarity 0.5774 is greater than the threshold. Therefore, the specific category of the target data under the indirect dimension data application field is determined to be the business environment. If the similarity threshold is set to 0.6, since the maximum similarity is less than 0.6, the specific category of the target data under the indirect dimension data application field is determined to be empty, and the specific category of the target data under this dimension can be filled manually later. According to the above method, the specific classification of the target data in all dimensions is obtained, and the classification process of the target data is completed. In the face of multi-dimensional classification needs. This method combines metadata analysis and multi-dimensional classification technology based on content analysis, significantly improves the efficiency and accuracy of data classification, reduces manual intervention, and optimizes the data management process.

[0101] S7. Determine the initial evaluation level of the target data based on the classification data of the target data.

[0102] The several levels from low to high are primary data, secondary data, tertiary data, quaternary data, important data and core data.

[0103] S8. Determine the level of the target data according to the initial evaluation level and rating process of the target data; store the classification data and level of the target data.

[0104] The step of determining the initial evaluation level of the target data according to the classification data of the target data is as follows:

[0105] Obtain the classification data of the target data, and obtain all specific categories in the classification data;

[0106] For each specific category, obtain the highest level among all levels of the stored public data of the specific category as the corresponding level of the specific category;

[0107] The highest level among the corresponding levels of all specific categories is obtained as the initial evaluation level of the target data.

[0108] In this embodiment, first, we obtain the classification data of the target data. For example, the classification data is system generated and business environment; next, we need to determine the initial evaluation level of the target data. To this end, we need to find the level of the stored public data with the same classification. Among the stored public data under the system generated category, the highest level is important data. Among the stored public data under the business environment category, the highest level is level three data. The highest level of these two levels is selected as the initial evaluation level of the target data. Therefore, the initial evaluation level of the target data is determined to be important data. Next, we determine its final level based on the initial evaluation level of the target data and the preset rating process to complete the data grading work.

[0109] Through this method, we can preliminarily determine the level of new data based on the classification information of the data and the level of historical data, and then verify and adjust this level through further evaluation processes. This improves the efficiency of data rating.

[0110] Example 2

[0111] Based on Example 1, the rating process includes:

[0112] Set a corresponding judgment indicator for each level;

[0113] Starting from the judgment indicator corresponding to the highest level, determine whether the data to be rated meets the judgment indicator of the current level;

[0114] If the judgment result is yes, the level of the data to be rated is determined to be the level corresponding to the current judgment indicator; if the judgment result is no, continue to judge the next lower level judgment indicator; until the level of the data to be rated is determined.

[0115] Determining the level of the target data according to the initial evaluation level and the rating process of the target data includes: obtaining the initial evaluation level, starting the rating process with the judgment index corresponding to the initial evaluation level, and determining the level of the target data.

[0116] In this embodiment, the rating system may include the following levels, from high to low, core data, important data, level 4 data, level 3 data, level 2 data, and level 1 data. Each level has a corresponding judgment indicator, and the judgment indicator of each level can be set accordingly according to the actual classification requirements. For example, the judgment indicator of level 4 data can be set as "can cause particularly serious harm to personal rights and organizational rights, or can cause general harm to public interests and social stability"; and the judgment indicator of level 3 data can be set as "can cause serious harm to personal rights and organizational rights". When rating, rating is carried out from high level to low level step by step to ensure accurate classification; in this embodiment, rating is carried out step by step from the initial evaluation level, which can effectively reduce the evaluation process and improve evaluation efficiency.

[0117] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only one, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0118] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

[0119] Finally: 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 spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-dimensional public data classification method, characterized in that: include: S1. Set multiple classification dimensions and specific categories under each classification dimension, set several levels from low to high, and establish a rating process; S2. Establish a mapping relationship between the metadata of public data and specific categories under the classification dimension, and mark the classification dimension as a direct dimension or an indirect dimension based on the mapping relationship; The marking of the classification dimension as a direct dimension or an indirect dimension based on the mapping relationship includes: marking the classification dimension of a specific category having a mapping relationship with the metadata as a direct dimension, and marking the classification dimension of a specific category having no mapping relationship with the metadata as an indirect dimension; S3, setting a corresponding keyword group for each indirect dimension, obtaining public data that has been classified in the past, and determining the classification vector of each specific category under the indirect dimension according to the set keyword group and the public data that has been classified in the past; The step of determining the classification vector of each specific category under the indirect dimension according to the set keyword group and the public data that has been classified in history includes: S31, obtaining any specific category under the indirect dimension as the analysis category, and obtaining a keyword group corresponding to the indirect dimension as the analysis phrase; S32. extract at least two pieces of public data classified as analysis categories from the public data that have been classified in the past; S33, performing feature extraction operations on each piece of extracted public data according to the analysis phrase, and obtaining a feature vector corresponding to each piece of public data; S34, obtaining a classification vector of the analysis category according to all feature vectors; S35, repeating steps S31 to S34 to obtain the classification vector of each specific category under the indirect dimension; The step of performing feature extraction operation comprises: Obtain keyword groups and public data, and obtain the number n of keywords in the keyword group; Create an array with n elements and record it as the characteristic array; Each element in the feature array is assigned a value using the following formula: Where, T i represents the i-th element in the feature array, G i represents the number of times the i-th keyword in the keyword group appears in the public data, G o Indicates the number of times the oth keyword in the keyword group appears in the public data, where o is the index of the keyword; Use the assigned feature array as the feature vector to complete the feature extraction operation; The step of obtaining the classification vector of the analysis category according to all the feature vectors comprises: Obtain all the feature vectors of the analysis category to form a feature vector set; establish a selection vector with the same number of feature vector elements; Synchronously traverse each feature vector and selection vector in the feature vector set in order, and extract the elements of all feature vectors at the traversed position during the traversal process; Calculate the median of the extracted elements, and use the calculated median as the element value of the selection vector at the traversal position; The selected vector array after traversal is used as the classification vector of the analysis category; S4. Obtain the public data to be classified and recorded as target data, obtain metadata of the target data, and obtain specific categories under the direct dimension according to the mapping relationship between the metadata and the category; S5. Perform feature extraction on the target data according to the keyword group of the indirect dimension to obtain a feature vector of the target data, and determine the specific category of the target data under the indirect dimension according to the feature vector of the target data and the feature vector of each category under the indirect dimension; S6. Use the specific categories of the target data under all direct dimensions and the specific categories of the target data under all indirect dimensions to form the classification data of the target data; S7. Determine the initial evaluation level of the target data according to the classification data of the target data; S8. Determine the level of the target data according to the initial evaluation level and rating process of the target data; store the classification data and level of the target data.

2. The multi-dimensional public data classification method according to claim 1 is characterized in that: The step of determining the specific category of the target data under the indirect dimension according to the feature vector of the target data and the feature vector of each classification under the indirect dimension comprises: Get the feature vector of the target data and all classification vectors under the indirect dimension; Calculate the similarity between the feature vector of the target data and each classification vector under the indirect dimension respectively; The calculation formula of the similarity is as follows: In the formula, XS represents similarity, a represents the feature vector of the target data, and b represents the classification vector; · represents the dot product operation, ||a|| represents the modulus length of the feature vector of the target data, ||b|| represents the modulus length of the classification vector, and a j represents the jth element value in the feature vector of the target data, b j represents the jth element value in the classification vector, and m represents the total number of elements in the classification vector; Obtaining the maximum value of all calculated similarities, and determining whether the maximum value is greater than or equal to a preset similarity threshold; If so, the specific category corresponding to the classification vector corresponding to the maximum value is taken as the specific category of the target data under the indirect dimension; If not, the specific category of the target data under the indirect dimension is set to empty.

3. The multi-dimensional public data classification method according to claim 2 is characterized in that: The several levels from low to high are primary data, secondary data, tertiary data, quaternary data, important data and core data.

4. The multi-dimensional public data classification method according to claim 3 is characterized in that: The step of determining the initial evaluation level of the target data according to the classification data of the target data is as follows: Get the classification data of the target data, get all the specific categories in the classification data, For each specific category, obtain the highest level among all levels of the stored public data of the specific category as the corresponding level of the specific category; The highest level among the corresponding levels of all specific categories is obtained as the initial evaluation level of the target data.

5. The multi-dimensional public data classification method according to claim 4 is characterized in that: The rating process includes: Set a corresponding judgment indicator for each level; Starting from the judgment indicator corresponding to the highest level, determine whether the data to be rated meets the judgment indicator of the current level; If the judgment result is yes, the level of the data to be rated is determined to be the level corresponding to the current judgment indicator; if the judgment result is no, continue to judge the next lower level judgment indicator; until the level of the data to be rated is determined.

6. A multi-dimensional public data classification method according to claim 5, characterized in that: Determining the level of the target data according to the initial evaluation level and the rating process of the target data includes: obtaining the initial evaluation level, starting the rating process with the judgment index corresponding to the initial evaluation level, and determining the level of the target data.

Images