An evaluation management method and system based on multi-dimensional data attributes

Abstract

The application discloses a kind of based on the evaluation management method and system of multidimensional data attribute, it is related to data processing technical field, the method includes: receiving the metadata information of target database carries out attribute extraction, obtains multiple metadata attribute sets;Multiple metadata attribute sets are input into cloud occupation degree space, obtain multiple occupation degree results;Traversal multiple occupation degree analysis results are carried out negative constraint, obtain multiple constraint metadata attribute sets;Multiple data source feature sets are constructed;Multiple data source feature sets are decomposed using T-SVD method, obtain multiple singular value sets;Filter set in accordance with preset singular value, mapping obtains multiple re-identification data source feature sets;Target database management is carried out.The present application solves the technical problems of low database management efficiency and unreasonable resource allocation in the prior art, and achieves the technical effect of improving the rationalization degree of database management.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and specifically to an evaluation and management method and system based on multidimensional data attributes. Background Technology

[0002] With the rapid development of information technology, the technical means of information recording are also constantly changing, from the initial paper-based document recording to electronic archiving. By summarizing electronically archived data into databases, information access becomes easier. Therefore, studying database management is of great significance for improving office efficiency.

[0003] Currently, due to the large variety and quantity of data within databases, uniform resource allocation results in slow data retrieval speeds, long response times to requests, and an inability to respond promptly to demands. Existing technologies suffer from low database management efficiency and unreasonable resource allocation. Summary of the Invention

[0004] This application provides an evaluation and management method and system based on multidimensional data attributes, which is used to address the technical problems of low database management efficiency and unreasonable resource allocation in the prior art.

[0005] In view of the above problems, this application provides an evaluation and management method and system based on multidimensional data attributes.

[0006] The first aspect of this application provides an evaluation and management method based on multidimensional data attributes, the method comprising:

[0007] Receive metadata information from the target database, extract attributes to obtain multiple metadata attribute sets, wherein each of the multiple metadata attribute sets corresponds one-to-one with the metadata information;

[0008] Input the multiple sets of metadata attributes into the cloud occupancy space to obtain multiple occupancy results, wherein the occupancy results include frequency-dimensional occupancy and time-dimensional occupancy;

[0009] By iterating through the multiple occupancy analysis results and applying negative constraints, a set of multiple constraint metadata attributes is obtained.

[0010] Multiple data source feature sets are constructed based on the aforementioned multiple constraint metadata attribute sets, wherein the data source is a data object described by metadata information;

[0011] The T-SVD method is used to decompose the feature sets of multiple data sources to obtain multiple sets of singular values;

[0012] The multiple sets of singular values ​​are traversed to filter the sets that match the preset singular values, and multiple sets of re-identification data source features are mapped to obtain the re-identification data source feature sets.

[0013] The target database is managed based on the feature sets of the multiple re-identification data sources.

[0014] A second aspect of this application provides an evaluation and management system based on multidimensional data attributes, the system comprising:

[0015] An attribute set acquisition module is used to receive metadata information from a target database, extract attributes, and obtain multiple metadata attribute sets, wherein the multiple metadata attribute sets correspond one-to-one with the metadata information;

[0016] The occupancy result acquisition module is used to input the multiple metadata attribute sets into the cloud occupancy space to obtain multiple occupancy results, wherein the occupancy results include frequency-dimensional occupancy and time-dimensional occupancy;

[0017] A constraint attribute set acquisition module is used to traverse the multiple occupancy analysis results to perform negative constraints and obtain multiple constraint metadata attribute sets.

[0018] The data source feature acquisition module is used to construct multiple data source feature sets based on the multiple constraint metadata attribute sets, wherein the data source is a data object described by metadata information;

[0019] A singular value acquisition module is used to decompose multiple data source feature sets using the T-SVD method to obtain multiple singular value sets.

[0020] The re-identification data source acquisition module is used to traverse the multiple singular value sets, filter the sets that meet the preset singular values, and map to obtain multiple re-identification data source feature sets.

[0021] A database management module is used to manage the target database based on the feature sets of the multiple re-identification data sources.

[0022] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0023] This application extracts attributes from the metadata information of the target database to obtain multiple metadata attribute sets, each corresponding one-to-one with the metadata information. These metadata attribute sets are then input into a cloud occupancy space to obtain multiple occupancy results, including frequency-dimensional occupancy and time-dimensional occupancy. These results are then iterated through and negative constraints are applied to obtain multiple constraint metadata attribute sets. Multiple data source feature sets are constructed based on these constraint metadata attribute sets, where the data source is a data object described by the metadata information. The T-SVD method is then used to decompose these multiple data source feature sets, resulting in multiple singular value sets. These singular value sets are then iterated through and sets matching preset singular values ​​are selected, mapping to multiple re-identified data source feature sets. Finally, target database management is performed based on these re-identified data source feature sets. This achieves the technical effect of improving data management quality and efficiency. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A schematic diagram of an evaluation and management method based on multidimensional data attributes provided in this application embodiment;

[0026] Figure 2 A flowchart illustrating the process of obtaining multiple occupancy results in an evaluation and management method based on multidimensional data attributes, provided in an embodiment of this application.

[0027] Figure 3 A schematic diagram illustrating the process of obtaining multiple constraint metadata attribute sets in an evaluation and management method based on multidimensional data attributes provided in this application embodiment;

[0028] Figure 4 This is a schematic diagram of an evaluation and management system structure based on multidimensional data attributes, provided as an embodiment of this application.

[0029] Attached diagram labels: Module 11 for obtaining attribute set, Module 12 for obtaining occupancy result, Module 13 for obtaining constraint attribute set, Module 14 for obtaining data source feature, Module 15 for obtaining singular value, Module 16 for re-identifying data source, and Module 17 for database management. Detailed Implementation

[0030] This application provides an evaluation and management method based on multidimensional data attributes to address the technical problems of low database management efficiency and unreasonable resource allocation in existing technologies.

[0031] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0032] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0033] Example 1

[0034] like Figure 1 As shown, this application provides an evaluation and management method based on multidimensional data attributes, the method comprising:

[0035] Step S100: Receive metadata information from the target database and extract attributes to obtain multiple metadata attribute sets, wherein each of the multiple metadata attribute sets corresponds one-to-one with the metadata information;

[0036] Specifically, the target database is any database used for data management. By extracting attributes from the database data, foundational analytical data is provided for subsequent data analysis. The multiple metadata information consists of multiple data information describing the type or function of the data, including structured metadata, descriptive metadata, management metadata, business metadata, technical metadata, and operational metadata. The multiple metadata attribute set is obtained by extracting the type attributes from the multiple metadata information, including structural attributes, descriptive attributes, management attributes, business attributes, technical attributes, and operational attributes. Since the multiple metadata attribute set is obtained by extracting attributes from the multiple metadata information, the attributes in the multiple metadata attribute set correspond one-to-one with the metadata information.

[0037] Step S200: Input the multiple metadata attribute sets into the cloud occupancy space to obtain multiple occupancy results, wherein the occupancy results include frequency-dimensional occupancy and time-dimensional occupancy;

[0038] Furthermore, such as Figure 2 As shown, the step S200 of this application embodiment further includes inputting the multiple metadata attribute sets into the cloud occupancy space to obtain multiple occupancy results:

[0039] Step S210: Obtain historical call information of the target database and construct a historical call information set, wherein the historical call information set includes multiple historical metadata attribute sets, multiple historical frequency dimension occupancy and multiple historical time dimension occupancy;

[0040] Step S220: Use the multiple historical metadata attribute sets and the multiple historical frequency dimension occupancy as the first construction data set to construct a frequency dimension decision forest;

[0041] Step S230: Use the multiple historical metadata attribute sets and the multiple historical time dimension occupancy as the second construction data set to construct a time dimension decision forest;

[0042] Step S240: Connect the frequency-dimensional decision forest and the time-dimensional decision forest in parallel to obtain the cloud occupancy space.

[0043] Specifically, the cloud occupancy space is a space for intelligent calculation of the occupancy of various metadata attributes in the database, including an input layer, a frequency-dimensional decision forest, a time-dimensional decision forest, and an output layer. The cloud occupancy space is constructed by connecting the frequency-dimensional and time-dimensional decision forests in parallel, then connecting the input layer before the parallel structure and the output layer after the parallel structure. The input data consists of multiple sets of metadata attributes, and the output data consists of multiple occupancy results. These multiple occupancy results describe the occupancy of each metadata attribute in the multiple metadata attribute sets in the database, including frequency-dimensional occupancy and time-dimensional occupancy. The frequency-dimensional occupancy describes the occupancy based on the frequency of attribute usage. The time-dimensional occupancy describes the occupancy based on the usage of the attribute over time.

[0044] Specifically, historical call data of the target database is extracted to obtain the historical call information set. This historical call information set describes the call information of the target database within a historical time period, including multiple historical metadata attribute sets, multiple historical frequency dimension occupancy values, and multiple historical time dimension occupancy values. The multiple historical metadata attribute sets and the multiple historical frequency dimension occupancy values ​​are used as the first construction data set to construct a frequency-dimensional decision forest. The first construction data set consists of data related to data frequency within the historical time period, which can be used to train and construct the frequency-dimensional decision forest, and includes multiple historical metadata attribute sets and the multiple historical frequency dimension occupancy values. The multiple historical metadata attribute sets and the multiple historical time dimension occupancy values ​​are used as the second construction data set to construct a time-dimensional decision forest. The second construction data set consists of data related to data usage time within the historical time period, which can be used to train and construct the time-dimensional decision forest, and includes multiple historical metadata attribute sets and the multiple historical time dimension occupancy values.

[0045] Specifically, the frequency-dimensional decision forest is an intelligent model that analyzes frequency from different perspectives, including a retrieval frequency decision tree, a storage frequency decision tree, and a labeling frequency decision tree. The retrieval frequency decision tree is a functional decision tree that intelligently analyzes the frequency of different metadata attributes being retrieved in the database. The storage frequency decision tree is a functional decision tree that intelligently analyzes the frequency of metadata attributes corresponding to data stored in the database. The labeling frequency decision tree is a decision tree that intelligently analyzes the number of times different metadata attributes are labeled by users; even if a metadata attribute is not used frequently enough, it still needs to be considered. The time-dimensional decision forest is an intelligent functional model that analyzes time from different perspectives, including a duration decision tree and a time period decision tree. The duration decision tree analyzes the time length corresponding to the metadata attribute, and the time period decision tree analyzes the time period with the highest retrieval frequency of the metadata attribute. By connecting the frequency-dimensional decision forest and the time-dimensional decision forest in parallel, the cloud occupancy space is obtained.

[0046] Furthermore, in the embodiment of this application, step S220 further includes using the multiple historical metadata attribute sets and the multiple historical frequency dimension occupancy as the first construction data set to construct a frequency dimension decision forest:

[0047] Step S221: Perform data association extraction on the multiple historical metadata attribute sets, and extract multiple attribute retrieval frequencies, multiple attribute storage frequencies, and multiple attribute tagging frequencies from the working logs of the target database;

[0048] Step S222: Construct a retrieval frequency decision tree based on the multiple historical metadata attribute sets and the multiple attribute retrieval frequencies;

[0049] Step S223: Construct a storage frequency decision tree based on the multiple historical metadata attribute sets and the multiple attribute storage frequencies;

[0050] Step S224: Construct a labeling frequency decision tree based on the multiple historical metadata attribute sets and the multiple attribute labeling frequencies;

[0051] Step S225: Construct the frequency-dimensional decision forest based on the retrieved frequency decision tree, the stored frequency decision tree, and the labeled frequency decision tree.

[0052] Furthermore, in this embodiment, step S225 of constructing the frequency-dimensional decision forest based on the retrieved frequency decision tree, the stored frequency decision tree, and the labeled frequency decision tree also includes:

[0053] Step S2251: Train and validate the weighted calculation layer using the multiple attribute retrieval frequency, the multiple attribute storage frequency, the multiple attribute labeling frequency, and the multiple historical frequency dimension occupancy to obtain the weighted calculation layer;

[0054] Step S2252: Embed the weighted calculation layer in the frequency dimension decision forest and interact with the retrieved frequency decision tree, the stored frequency decision tree, and the labeled frequency decision tree.

[0055] Specifically, the work log is the work record data during the use of the target database. By extracting data from the work log, the retrieval frequency, storage frequency, and tagging frequency of multiple attributes can be obtained. The retrieval frequency of multiple attributes is obtained by statistically analyzing the number of times each historical metadata attribute is retrieved. The storage frequency of multiple attributes is obtained by statistically analyzing the number of times each historical metadata attribute is stored. The tagging frequency of multiple attributes is obtained by statistically analyzing the number of times each historical metadata attribute is tagged by the user.

[0056] Specifically, an attribute storage frequency is randomly selected from the multiple attribute storage frequencies as a first partitioning node. The multiple attribute storage frequencies are then binary-classified to obtain a first partitioning result. Next, an attribute storage frequency is randomly selected from the multiple attribute storage frequencies as a second partitioning node. The attribute storage frequencies corresponding to the first and second partitioning nodes are not the same. The multiple attribute storage frequencies are then binary-classified to obtain a second partitioning result. This process is repeated to construct multiple partitioning nodes to obtain a final partitioning result. Then, based on the one-to-one correspondence between the attribute storage frequencies in the partitioning results and the historical metadata attribute set, the partitioning results are labeled according to the historical metadata attribute set. Using the multiple historical metadata attribute sets as the root node and the multiple partitioning nodes as leaf nodes, the retrieval frequency decision tree is obtained.

[0057] Specifically, an attribute storage frequency is randomly selected from the multiple attribute storage frequencies as a first partitioning node. The multiple attribute storage frequencies are then binary-classified to obtain a first partitioning result. Next, an attribute storage frequency is randomly selected from the multiple attribute storage frequencies as a second partitioning node. The attribute storage frequencies corresponding to the first and second partitioning nodes are not the same. The multiple attribute storage frequencies are then binary-classified to obtain a second partitioning result. This process is repeated to construct multiple partitioning nodes to obtain a final partitioning result. Then, based on the one-to-one correspondence between the attribute storage frequencies in the partitioning results and the historical metadata attribute set, the partitioning results are labeled according to the historical metadata attribute set. Using the multiple historical metadata attribute sets as the root node and the multiple partitioning nodes as leaf nodes, the storage frequency decision tree is obtained.

[0058] Specifically, an attribute label frequency is randomly selected from the multiple attribute label frequencies as a first splitting node. The multiple attribute label frequencies are then binary-classified to obtain a first splitting result. Next, an attribute label frequency is randomly selected from the multiple attribute label frequencies as a second splitting node. The attribute label frequencies corresponding to the first splitting node and the second splitting node are not the same. The multiple attribute label frequencies are then binary-classified to obtain a second splitting result. This process of constructing multiple splitting nodes is repeated multiple times to obtain a final splitting result. Then, based on the one-to-one correspondence between the attribute label frequencies in the splitting results and the historical metadata attribute set, the splitting results are labeled according to the historical metadata attribute set. Using the multiple historical metadata attribute sets as the root node and the multiple splitting nodes as leaf nodes, the label frequency decision tree is obtained.

[0059] Specifically, the historical dataset is constructed by using the retrieval frequency, storage frequency, labeling frequency, and historical frequency dimension occupancy of multiple attributes as a historical dataset. This historical dataset is divided into a training set and a validation set according to a certain ratio, preferably 2:1. The weighted computation layer based on a backpropagation neural network is trained using the training set until convergence. Then, using the validation set, the retrieval frequency, storage frequency, and labeling frequency of multiple attributes are input into the weighted computation layer to obtain the frequency dimension occupancy. The frequency dimension occupancy is compared with the historical frequency dimension occupancy in the validation set. When the matching result reaches 60% or more, the accuracy meets the requirements, and the weighted computation layer is obtained. By embedding the weighted computation layer in the frequency-dimensional decision forest, weighted computation can be performed on the frequencies of the three dimensions.

[0060] Step S300: Traverse the multiple occupancy analysis results and apply negative constraints to obtain multiple constraint metadata attribute sets;

[0061] Furthermore, such as Figure 3 As shown, the step of traversing the multiple occupancy analysis results to perform negative constraints and obtain multiple constraint metadata attribute sets, step S300 of this application embodiment further includes:

[0062] Step S310: Traverse the multiple occupancy analysis results and extract multiple frequency-dimensional occupancy.

[0063] Step S320: Set a frequency dimension occupancy threshold, and use the frequency dimension occupancy threshold as the first negative constraint condition;

[0064] Step S330: Apply a negative constraint to the multiple occupancy analysis results according to the first negative constraint condition to obtain the first negative constraint result.

[0065] Furthermore, after obtaining the first negative constraint result, step S330 of this application embodiment further includes:

[0066] Step S331: Traverse the first negative constraint result and extract multiple time dimension occupancy, wherein the multiple time dimension occupancy corresponds one-to-one with the occupancy analysis result in the first negative constraint result;

[0067] Step S332: Obtain and parse user requirement information to obtain the negative constraint time period;

[0068] Step S333: Use the time period that satisfies the negation constraint as the second negation constraint condition;

[0069] Step S334: Apply a negative constraint to the first negative constraint result according to the second negative constraint condition to obtain the set of multiple constraint metadata attributes.

[0070] Specifically, applying negative constraints to the multiple occupancy analysis results refers to optimizing and filtering the multiple occupancy analysis results, performing a dimensionality reduction, and reducing the number of analyses. The multiple constraint metadata attribute set is the set of attributes worth analyzing after applying negative constraints. The frequency dimension occupancy is extracted from the multiple occupancy analysis results to obtain the multiple frequency dimension occupancy. The frequency dimension occupancy threshold is the lowest frequency value that meets the analysis requirements. The first negative constraint condition is the condition information for constraining the multiple occupancy analysis results from a frequency perspective. Results below the frequency dimension occupancy threshold are removed, and the remaining results are used as the first negative constraint results.

[0071] Specifically, the user requirement information describes the user's requirements for data usage, including data type and usage time. The negative constraint time is the data access time period determined based on the user's requirements. The second negative constraint condition is a condition that constrains multiple occupancy analysis results in terms of time. By filtering out occupancy analysis results in the first negative constraint results whose time periods are not within the range of the second negative constraint condition, the set of multiple constraint metadata attributes is obtained.

[0072] Step S400: Construct multiple data source feature sets based on the multiple constraint metadata attribute sets, wherein the data source is a data object described by metadata information;

[0073] Specifically, the data source is the data described by metadata information. By using the multiple sets of constraint metadata attributes and the correspondence between these sets and the metadata information, the constraint metadata attributes are used as features of the data source, resulting in the multiple sets of data source feature sets. These multiple sets of data source feature sets are collections that describe the data characteristics of the data source.

[0074] Step S500: Use the T-SVD method to decompose the feature sets of multiple data sources to obtain multiple sets of singular values;

[0075] Specifically, Tensor Singular Value Decomposition (T-SVD) is a novel decomposition method that can shift data from the time domain to the frequency domain using Fourier transform, effectively analyzing the characteristics of data in the frequency domain and enabling data compression. By using the T-SVD method to decompose feature sets from multiple data sources and calculating data features, multiple singular value sets are obtained. These multiple singular value sets are data values ​​calculated from the features of multiple data sources, and each singular value set corresponds one-to-one with a feature from a different data source.

[0076] Step S600: Traverse the multiple singular value sets to filter the sets that meet the preset singular values, and map them to obtain multiple re-identification data source feature sets;

[0077] Specifically, the preset singular values ​​are the minimum singular values ​​that meet user requirements. The multiple sets of singular values ​​are filtered, eliminating those below the preset singular values, resulting in multiple sets of singular values ​​that meet the requirements. Then, based on the one-to-one correspondence between the multiple sets of singular values ​​and multiple sets of data source features, the multiple sets of re-identified data source features are obtained, thereby performing a second dimensionality reduction on the data source feature sets. These multiple sets of re-identified data source features are obtained after data filtering from the perspective of singular values.

[0078] Step S700: Perform target database management based on the multiple re-identification data source feature sets.

[0079] Furthermore, in the step S700 of this application embodiment, the management of the target database based on the multiple re-identification data source feature sets further includes:

[0080] Step S710: Serialize and sort the multiple singular value sets corresponding to the multiple re-identification data source feature sets to obtain the data source sequence to be called;

[0081] Step S720: Obtain a computing power optimization scheme based on the sequence of data sources to be invoked;

[0082] Step S730: Manage the target database according to the computing power optimization scheme.

[0083] Specifically, the multiple singular value sets corresponding to the multiple re-identification data source feature sets are serialized and sorted in descending order of singular values ​​to obtain the sequence of data sources to be invoked. This sequence is a queue formed by arranging the data sources in the order they will be invoked. A computing power optimization scheme is obtained based on this sequence, where the scheme optimizes the computing power allocation of the database; data sources listed earlier receive more computing power. This achieves optimized management of the database.

[0084] In summary, the embodiments of this application have at least the following technical effects:

[0085] This application embodiment extracts attributes from the metadata information of the target database to obtain multiple metadata attribute sets, achieving the goal of metadata attribute extraction. Then, by inputting these multiple metadata attribute sets into a cloud occupancy space, multiple occupancy results are obtained, including frequency-dimensional occupancy and time-dimensional occupancy. Negative constraints are applied to these multiple occupancy analysis results along both frequency and time dimensions. Based on these constraints, multiple data source feature sets are constructed. The data source is a data object described by the metadata information. The T-SVD method is used to decompose these multiple data source feature sets, calculating singular values ​​to provide a basis for subsequent secondary dimensionality reduction of the data source features. Multiple singular value sets are traversed to filter sets that meet preset singular values, mapping to obtain multiple re-identified data source feature sets. The target database is then managed based on these multiple re-identified data source feature sets. This achieves the technical effect of efficient database management and improved management efficiency.

[0086] Example 2

[0087] Based on the same inventive concept as the assessment and management method based on multidimensional data attributes in the foregoing embodiments, such as Figure 4 As shown, this application provides an evaluation and management system based on multi-dimensional data attributes. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0088] The attribute set acquisition module 11 is used to receive metadata information from the target database, extract attributes, and obtain multiple metadata attribute sets, wherein the multiple metadata attribute sets correspond one-to-one with the metadata information;

[0089] The occupancy result acquisition module 12 is used to input the multiple metadata attribute sets into the cloud occupancy space to obtain multiple occupancy results, wherein the occupancy results include frequency dimension occupancy and time dimension occupancy;

[0090] The constraint attribute set acquisition module 13 is used to traverse the multiple occupancy analysis results to perform negative constraints and obtain multiple constraint metadata attribute sets.

[0091] The data source feature acquisition module 14 is used to construct multiple data source feature sets based on the multiple constraint metadata attribute sets, wherein the data source is a data object described by metadata information;

[0092] Singular value acquisition module 15, which is used to decompose multiple data source feature sets using the T-SVD method to obtain multiple singular value sets;

[0093] The re-identification data source acquisition module 16 is used to traverse the multiple singular value sets, filter the sets that meet the preset singular values, and map to obtain multiple re-identification data source feature sets.

[0094] Database management module 17 is used to manage the target database based on the multiple re-identification data source feature sets.

[0095] Furthermore, the system also includes:

[0096] A call information set acquisition unit is used to acquire historical call information of the target database and construct a historical call information set, wherein the historical call information set includes multiple historical metadata attribute sets, multiple historical frequency dimension occupancy and multiple historical time dimension occupancy;

[0097] A frequency-dimensional decision forest construction unit is used to construct a frequency-dimensional decision forest by taking the multiple historical metadata attribute sets and the multiple historical frequency dimension occupancy as the first construction data set.

[0098] A time-dimensional decision forest construction unit is used to construct a time-dimensional decision forest by taking the multiple sets of historical metadata attributes and the multiple sets of historical time dimension occupancy as a second construction data set.

[0099] A cloud occupancy space acquisition unit is used to connect the frequency-dimensional decision forest and the time-dimensional decision forest in parallel to obtain the cloud occupancy space.

[0100] Furthermore, the system also includes:

[0101] A frequency extraction unit is used to perform data association extraction on the multiple historical metadata attribute sets, extract multiple attribute retrieval frequencies, multiple attribute storage frequencies, and multiple attribute tagging frequencies from the working logs of the target database.

[0102] A retrieval frequency decision tree construction unit is used to construct a retrieval frequency decision tree based on the multiple historical metadata attribute sets and the multiple attribute retrieval frequencies.

[0103] A storage frequency decision tree construction unit is used to construct a storage frequency decision tree based on the multiple historical metadata attribute sets and the multiple attribute storage frequencies.

[0104] A labeling frequency decision tree construction unit is used to construct a labeling frequency decision tree based on the plurality of historical metadata attribute sets and the plurality of attribute labeling frequencies;

[0105] A decision forest construction unit is used to construct the frequency-dimensional decision forest based on the retrieved frequency decision tree, the stored frequency decision tree, and the labeled frequency decision tree.

[0106] Furthermore, the system also includes;

[0107] A weighted computation layer obtaining unit is used to train and verify the weighted computation layer using the multiple attribute retrieval frequency, the multiple attribute storage frequency, the multiple attribute labeling frequency, and the multiple historical frequency dimension occupancy, so as to obtain the weighted computation layer.

[0108] An interactive communication unit is used to embed the weighted calculation layer into the frequency-dimensional decision forest and to interact and communicate with the retrieved frequency decision tree, the stored frequency decision tree, and the labeled frequency decision tree.

[0109] Furthermore, the system also includes:

[0110] Multiple frequency dimension occupancy extraction units are used to traverse the multiple occupancy analysis results and extract multiple frequency dimension occupancy.

[0111] The first negative constraint condition obtaining unit is used to set a frequency dimension occupancy threshold and use the frequency dimension occupancy threshold as the first negative constraint condition.

[0112] The first negative constraint result obtaining unit is used to apply negative constraints to the multiple occupancy analysis results according to the first negative constraint condition to obtain the first negative constraint result.

[0113] Furthermore, the system also includes:

[0114] Multiple time dimension occupancy extraction units are used to traverse the first negative constraint result and extract multiple time dimension occupancy, wherein the multiple time dimension occupancy corresponds one-to-one with the occupancy analysis result in the first negative constraint result;

[0115] A negative constraint time period acquisition unit is used to acquire user demand information and parse it to obtain the negative constraint time period.

[0116] The second negative constraint setting unit is used to set the time period that satisfies the negative constraint as the second negative constraint.

[0117] Multiple constraint metadata acquisition units are configured to perform negative constraints on the first negative constraint result according to the second negative constraint condition, thereby obtaining the multiple constraint metadata attribute sets.

[0118] Furthermore, the system also includes:

[0119] A data source sequence acquisition unit is used to perform serialization sorting based on multiple singular value sets corresponding to the multiple re-identification data source feature sets to obtain a data source sequence to be invoked.

[0120] A computing power optimization scheme acquisition unit is used to obtain a computing power optimization scheme based on the sequence of data sources to be invoked.

[0121] A target database management unit is used to manage the target database according to the computing power optimization scheme.

[0122] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, specific embodiments have been described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims can be performed in a different order than that shown in the embodiments and still achieve the desired result. Additionally, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0123] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0124] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. An evaluation and management method based on multidimensional data attributes, characterized in that, The method includes: Receive metadata information from the target database, extract attributes to obtain multiple metadata attribute sets, wherein each of the multiple metadata attribute sets corresponds one-to-one with the metadata information; Input the multiple sets of metadata attributes into the cloud occupancy space to obtain multiple occupancy results, wherein the occupancy results include frequency-dimensional occupancy and time-dimensional occupancy; By iterating through the multiple occupancy results and applying negative constraints, a set of multiple constraint metadata attributes is obtained; Multiple data source feature sets are constructed based on the aforementioned multiple constraint metadata attribute sets, wherein the data source is a data object described by metadata information; The T-SVD method is used to decompose the feature sets of multiple data sources to obtain multiple sets of singular values; The multiple sets of singular values ​​are traversed to filter the sets that match the preset singular values, and multiple sets of re-identification data source features are mapped to obtain the re-identification data source feature sets. The target database is managed based on the multiple re-identification data source feature sets; The step of inputting the multiple metadata attribute sets into the cloud occupancy space to obtain multiple occupancy results includes: Obtain historical call information of the target database and construct a historical call information set, wherein the historical call information set includes multiple historical metadata attribute sets, multiple historical frequency dimension occupancy and multiple historical time dimension occupancy; Using the multiple sets of historical metadata attributes and the multiple sets of historical frequency dimension occupancy as the first construction data set, a frequency dimension decision forest is constructed. The multiple sets of historical metadata attributes and the multiple historical time dimension occupancy are used as the second construction data set to construct a time dimension decision forest. By connecting the frequency-dimensional decision forest and the time-dimensional decision forest in parallel, the cloud occupancy space is obtained.

2. The method as described in claim 1, characterized in that, The step of constructing a frequency-dimensional decision forest by using the multiple sets of historical metadata attributes and the multiple historical frequency dimension occupancy as the first construction data set includes: Data association and extraction are performed on the multiple historical metadata attribute sets, and multiple attribute retrieval frequencies, multiple attribute storage frequencies, and multiple attribute tagging frequencies are extracted from the working logs of the target database. Based on the multiple sets of historical metadata attributes and the multiple attribute retrieval frequencies, a retrieval frequency decision tree is constructed. Construct a storage frequency decision tree based on the multiple sets of historical metadata attributes and the storage frequency of the multiple attributes; Construct a labeling frequency decision tree based on the multiple sets of historical metadata attributes and the multiple attribute labeling frequencies; The frequency-dimensional decision forest is constructed based on the retrieved frequency decision tree, the stored frequency decision tree, and the labeled frequency decision tree.

3. The method as described in claim 2, characterized in that, The method of constructing the frequency-dimensional decision forest based on the retrieved frequency decision tree, the stored frequency decision tree, and the labeled frequency decision tree further includes: The weighted computation layer is trained and validated by the retrieval frequency of the multiple attributes, the storage frequency of the multiple attributes, the labeling frequency of the multiple attributes, and the occupancy of the multiple historical frequency dimensions to obtain the weighted computation layer; The weighted calculation layer is embedded in the frequency-dimensional decision forest and interacts and communicates with the retrieved frequency decision tree, the stored frequency decision tree, and the labeled frequency decision tree.

4. The method as described in claim 1, characterized in that, The process of traversing the multiple occupancy results and applying negative constraints yields a set of multiple constraint metadata attributes, including: Traverse the multiple occupancy results and extract occupancy scores in multiple frequency dimensions; Set a frequency dimension occupancy threshold, and use the satisfaction of the frequency dimension occupancy threshold as the first negative constraint condition; The first negative constraint result is obtained by applying a negative constraint to the multiple occupancy results based on the first negative constraint condition.

5. The method as described in claim 4, characterized in that, After obtaining the first negative constraint result, the process includes: Traverse the first negative constraint result and extract multiple time dimension occupancy values, wherein the multiple time dimension occupancy values ​​correspond one-to-one with the occupancy value results in the first negative constraint result; Obtain and parse user demand information to obtain the negative constraint time period; The time period that satisfies the aforementioned negation constraint is taken as the second negation constraint condition; The first negative constraint result is negatively constrained according to the second negative constraint condition to obtain the set of multiple constraint metadata attributes.

6. The method as described in claim 1, characterized in that, The step of managing the target database based on the feature sets of the multiple re-identification data sources includes: The multiple singular value sets corresponding to the multiple re-identification data source feature sets are serialized and sorted to obtain the data source sequence to be invoked; A computing power optimization scheme is obtained based on the sequence of data sources to be invoked; The target database is managed according to the computing power optimization scheme.

7. An evaluation and management system based on multidimensional data attributes, characterized in that, The system includes: An attribute set acquisition module is used to receive metadata information from a target database, extract attributes, and obtain multiple metadata attribute sets, wherein the multiple metadata attribute sets correspond one-to-one with the metadata information; The occupancy result acquisition module is used to input the multiple metadata attribute sets into the cloud occupancy space to obtain multiple occupancy results, wherein the occupancy results include frequency-dimensional occupancy and time-dimensional occupancy; A constraint attribute set acquisition module is used to traverse the multiple occupancy results to perform negative constraints and obtain multiple constraint metadata attribute sets. The data source feature acquisition module is used to construct multiple data source feature sets based on the multiple constraint metadata attribute sets, wherein the data source is a data object described by metadata information; A singular value acquisition module is used to decompose multiple data source feature sets using the T-SVD method to obtain multiple singular value sets. The re-identification data source acquisition module is used to traverse the multiple singular value sets, filter the sets that meet the preset singular values, and map to obtain multiple re-identification data source feature sets. The database management module is used to manage the target database based on the multiple re-identification data source feature sets; The step of inputting the multiple metadata attribute sets into the cloud occupancy space to obtain multiple occupancy results includes: Obtain historical call information of the target database and construct a historical call information set, wherein the historical call information set includes multiple historical metadata attribute sets, multiple historical frequency dimension occupancy and multiple historical time dimension occupancy; Using the multiple sets of historical metadata attributes and the multiple sets of historical frequency dimension occupancy as the first construction data set, a frequency dimension decision forest is constructed. The multiple sets of historical metadata attributes and the multiple historical time dimension occupancy are used as the second construction data set to construct a time dimension decision forest. By connecting the frequency-dimensional decision forest and the time-dimensional decision forest in parallel, the cloud occupancy space is obtained.

Images