A data marking method and system based on database intelligent management

Abstract

The present invention relates to the technical field of database intelligent management, and discloses a data labeling method and system based on database intelligent management, including: data collection, including obtaining original data from multiple data sources; assigning data features to the original data, and generating variation features based on the original data; performing feature reconstruction according to the variation situation, and generating the labeling results of the original data; storing the original data in a database, and updating the corresponding index information; when retrieving data, retrieving the index-labeled data at the same time through the index. It realizes efficient storage, retrieval and management of massive data, and overcomes the limitations of traditional data management methods in processing complex and heterogeneous data. It can accurately predict data change trends and equipment response situations, and improve the accuracy and reliability of data analysis. It realizes automatic labeling and index updating of data, reduces manual intervention, and improves the intelligence level and automation level of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of database intelligent management, and in particular to a data marking method and system based on database intelligent management. Background Art

[0002] With the advent of the information age, the computing needs that need to be processed are also growing exponentially. Faced with massive amounts of data, with the development of modern industry and information technology, the importance of data in various fields is increasing. Especially in critical infrastructure such as power plants, the collection, processing and management of data have a vital impact on operational efficiency and safety. A large amount of data is generated during the operation of power plants, and this data needs to be collected, stored and analyzed in real time to support decision-making and optimize control. However, when faced with massive data and complex multi-source data, traditional data management methods have problems such as decentralized data storage, low retrieval efficiency, and insufficient data analysis capabilities. Therefore, how to efficiently manage and utilize this data has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] In view of the above-mentioned problems, the present invention is proposed.

[0004] Therefore, the technical problem solved by the present invention is that the existing data labeling method has the problems of too simple labeling and too single impact of the data retrieved during query.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a data marking method based on database intelligent management, comprising:

[0006] Data collection, including obtaining raw data from multiple data sources;

[0007] Assigning data features to the original data, and generating variation features based on the original data;

[0008] According to the variation situation, feature reconstruction is performed to generate a labeling result of the original data;

[0009] The original data is stored in the database and the corresponding index information is updated; when retrieving data, the data marked by the index is retrieved at the same time through the index.

[0010] As a preferred solution of the data labeling method based on database intelligent management described in the present invention, wherein: the original data includes, when the data information generated in the power plant is input into the database, the data source information of the input data is captured at the same time;

[0011] The input data and its data source are regarded as raw data, and the input data are initially classified according to the data source;

[0012] Original data format:

[0013] ,

[0014] in, Represents the nth data in the i-th data source, Indicates the nth data source.

[0015] As a preferred solution of the data labeling method based on database intelligent management described in the present invention, the data features are assigned to the original data, including presetting a standard value for each type of data, combining the current value, and calculating the deviation:

[0016] ,

[0017] in, Indicates the jth data in the i-th data source, which is the current value; Indicates the preset standard value of the j-th data in the i-th data source; Indicates the standard value of the jth data in the i-th data source;

[0018] Assume that each data can cause the response of other data and other devices except itself; for the current data The standard value and its deviation, using neural network to predict other data responses and other equipment responses that can be caused;

[0019] Input Layer:

[0020] ,

[0021] Hidden layer calculation:

[0022] ,

[0023] Output layer calculation:

[0024] ,

[0025] ,

[0026] Combined eigenvectors:

[0027] ,

[0028] in, Represents the input of the neural network; Represents the output of the l-1th hidden layer of the network; Represents the weight matrix of the lth layer of the neural network; represents the bias vector of the lth layer of the neural network, and The weight matrices representing parameter changes and equipment changes respectively; and denote the bias vectors of parameter changes and device changes, respectively; f(...) denotes the activation function of ReLU; L denotes the total number of hidden layers; Represents the current value of the jth data item in the i-th data source, which is a data set that can trigger a response; Represents the current value of the jth data item in the i-th data source, and the set of devices that can cause a response; represents a feature vector.

[0029] As a preferred solution of the data marking method based on database intelligent management described in the present invention, the variation characteristics include: for each original data , based on the historical records of the same data, calculate the coefficient of variation of the data:

[0030] Historical average:

[0031] ,

[0032] Standard Deviation:

[0033] ,

[0034] Coefficient of Variation:

[0035] ,

[0036] For the original data To mutate:

[0037] d ij ， = d ij [1+rand （ 0, CV ij ） ] ,

[0038] in, Represents the original data after mutation, Indicates 0 to A random number between express The maximum value of the same type of data, t represents the index of the record, and T represents the number of records.

[0039] As a preferred solution of the data labeling method based on database intelligent management described in the present invention, the feature reconstruction includes recalculating the combined feature vector using a neural network to obtain a mutated combined feature vector , take the union of the combined feature vectors before and after the mutation to form the final combined feature vector:

[0040] ,

[0041] ,

[0042] ,

[0043] ,

[0044] in, Represents the current value of the jth data item in the i-th data source, and the final set of data that can trigger a response; Represents the current value of the jth data in the i-th data source, and the set of data mutations that can cause a response; Represents the current value of the jth data item in the i-th data source, and the final set of devices that can cause a response; Represents the current value of the jth data item in the i-th data source, and the set of device mutations that can cause a response.

[0045] As a preferred solution of the data labeling method based on database intelligent management described in the present invention, the labeling result includes: And the final collection of equipment Mark each element in the data source, the original data and the timestamp;

[0046] At the same time, index information containing the tag features is generated according to the tag, and the data source, original data and timestamp can be obtained through the index information.

[0047] As a preferred solution of the data labeling method based on database intelligent management described in the present invention, wherein: the updating of the corresponding index information includes, when each data is stored in the database, obtaining the final combined feature vector by calculating the original data, thereby constructing the index information of the original data; and updating the index information in each data in the database;

[0048] The method of simultaneously retrieving the index-tagged data through the index includes: when querying data A, data A itself has multiple index information, and capturing the tag features pointed to by the index; when outputting data A, the captured tag features are simultaneously retrieved and output together.

[0049] A data marking system based on database intelligent management using the method of the present invention, wherein:

[0050] Acquisition unit, data acquisition, including obtaining raw data from multiple data sources;

[0051] A processing unit assigns data features to the original data, and generates variation features according to the original data; performs feature reconstruction according to the variation, and generates a labeling result of the original data;

[0052] The query unit stores the original data in the database and updates the corresponding index information; when retrieving data, the data marked by the index is retrieved at the same time through the index.

[0053] A computer device comprises: a memory and a processor; the memory stores a computer program, wherein: the processor implements the steps of any one of the methods of the present invention when executing the computer program.

[0054] A computer-readable storage medium stores a computer program, wherein: when the computer program is executed by a processor, the steps of any one of the methods of the present invention are implemented.

[0055] Beneficial effects of the present invention: The data labeling method based on intelligent database management provided by the present invention realizes efficient storage, retrieval and management of massive data by intelligently processing and labeling multi-source data, overcoming the limitations of traditional data management methods in processing complex and heterogeneous data. Combining feature extraction, variation analysis and neural network models, it can accurately predict data change trends and equipment response conditions, improve the accuracy and reliability of data analysis, and provide a scientific basis for decision support. Through real-time data collection and labeling, it is possible to monitor equipment status in a timely manner, warn of faults, optimize operating parameters, and improve the operating efficiency and safety of the system. The use of intelligent algorithms and neural network technology realizes automatic labeling and index updates of data, reduces manual intervention, and improves the intelligence level and automation level of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use 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 ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0057] Figure 1 An overall flow chart of a data labeling method based on database intelligent management provided for the first embodiment of the present invention. DETAILED DESCRIPTION

[0058] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0059] Example 1, reference Figure 1 , as an embodiment of the present invention, provides a data marking method based on database intelligent management, comprising:

[0060] S1: Data collection, including obtaining raw data from multiple data sources.

[0061] When entering the data information generated in the power plant into the database, the data source information of the input data is captured at the same time. The input data and its data source are regarded as raw data, and the input data is initially classified according to the data source. Raw data form:

[0062] ,

[0063] in, Represents the nth data in the i-th data source, Indicates the nth data source.

[0064] It is important to note that when obtaining raw data from multiple data sources and entering the data information generated by the power plant into the database, capturing the data source information of the input data at the same time helps to integrate data from different sources. This not only allows comprehensive collection of various operating parameters and status information, but also enables tracking the source of the data, providing a basis for subsequent data processing, analysis, and traceability. Taking the input data and its data source as raw data and initially classifying the input data according to the data source helps to systematically organize and manage the data. Initial classification helps to improve the efficiency of data storage and retrieval, ensuring that data from a specific source can be quickly and accurately located and extracted when needed.

[0065] S2: assigning data features to the original data, and generating variation features based on the original data.

[0066] Furthermore, assigning data features to the original data includes presetting a standard value for each type of data, combining the current value, and calculating the deviation:

[0067] ,

[0068] in, Indicates the jth data in the i-th data source, which is the current value; Indicates the preset standard value of the j-th data in the i-th data source; Indicates the standard value of the j-th data in the i-th data source.

[0069] By presetting the standard value of each data and calculating the deviation in combination with the current value, the anomalies and changes in the data can be accurately identified. The anomalies in the data can be effectively identified and provide a basis for subsequent analysis. By capturing these deviations, changes and potential problems in the operation of the power plant can be better understood, providing important information for fault prediction and maintenance.

[0070] Assume that each data can cause the response of other data and other devices except itself; for the current data The standard values ​​and their deviations are used to predict other data responses and other device responses that can be caused.

[0071] Input Layer:

[0072] ,

[0073] Hidden layer calculation:

[0074] ,

[0075] Output layer calculation:

[0076] ,

[0077] ,

[0078] Combined eigenvectors:

[0079] ,

[0080] in, Represents the input of the neural network; Represents the output of the l-1th hidden layer of the network; Represents the weight matrix of the lth layer of the neural network; represents the bias vector of the lth layer of the neural network, and The weight matrices representing parameter changes and equipment changes respectively; and denote the bias vectors of parameter changes and device changes, respectively; f(...) denotes the activation function of ReLU; L denotes the total number of hidden layers; Represents the current value of the jth data item in the i-th data source, which is a data set that can trigger a response; Represents the current value of the jth data item in the i-th data source, and the set of devices that can cause a response; represents a feature vector.

[0081] What we want to say is that by using neural networks to predict its potential impact on other data and equipment, a comprehensive response prediction model can be formed. This method can help us predict the chain reactions that may occur in the system in advance and take timely measures to avoid the expansion and spread of faults.

[0082] Through the neural network model, we can handle complex nonlinear relationships and provide accurate prediction results. Specifically, the input layer of the neural network receives the current value, standard value and deviation, the hidden layer performs nonlinear transformation, and the output layer generates predicted values ​​for other data and equipment responses. This prediction can not only improve the system's early warning capabilities, but also optimize the operation strategy. For example, when a data item deviates from the standard value, the system can predict the change trend of other related data and the operating status of the equipment, so as to make adjustments and interventions in advance to ensure the safe and stable operation of the power plant. The purpose of this design is to enhance the system's adaptability and intelligence level, realize comprehensive and real-time monitoring and management of power plant data, and improve overall operating efficiency and safety.

[0083] Furthermore, the variation characteristics include, for each original data , based on the historical records of the same data, calculate the coefficient of variation of the data.

[0084] Historical average:

[0085] ,

[0086] Standard Deviation:

[0087] ,

[0088] Coefficient of Variation:

[0089] ,

[0090] For the original data To mutate:

[0091] d ij ， = d ij [1+rand （ 0, CV ij ） ] ,

[0092] in, Represents the original data after mutation, Indicates 0 to A random number between express The maximum value of the same type of data, t represents the index of the record, and T represents the number of records.

[0093] S3: Reconstruct features according to the variation and generate labeling results of the original data.

[0094] Use the neural network to recalculate the combined feature vector to obtain the mutated combined feature vector , take the union of the combined feature vectors before and after the mutation to form the final combined feature vector:

[0095] ,

[0096] ,

[0097] ,

[0098] ,

[0099] in, Represents the current value of the jth data item in the i-th data source, and the final set of data that can trigger a response; Represents the current value of the jth data in the i-th data source, and the set of data mutations that can cause a response; Represents the current value of the jth data item in the i-th data source, and the final set of devices that can cause a response; Represents the current value of the jth data item in the i-th data source, and the set of device mutations that can cause a response.

[0100] The combined feature vectors before and after the mutation are combined to form the final combined feature vector, which can comprehensively capture the response characteristics of data and equipment in different states and provide complete information for further analysis and decision-making.

[0101] The labeling results include: And the final collection of equipment Each element in is marked, and the marking features are: data source, original data and timestamp; at the same time, index information containing the marking features is generated according to the marking, and the data source, original data and timestamp can be obtained through the index information.

[0102] It should be said that by marking this information, including data source, original data and timestamp, comprehensive tracking and tracing of data and equipment status can be achieved. Such a design helps to quickly locate the source and scope of the problem when an abnormal situation occurs, and improve the efficiency of fault handling and recovery. In addition, by marking to generate index information containing marking features, rapid retrieval and management of data and equipment status can be achieved. The update and maintenance of index information ensures the consistency and accuracy of data in the database, and supports real-time monitoring and intelligent analysis of the system. This design not only enhances the intelligence level of the system, but also improves the accuracy and reliability of data management, providing a strong guarantee for the stable operation and optimization of the power plant.

[0103] S4: The original data is stored in the database, and the corresponding index information is updated; when the data is retrieved, the data marked by the index is retrieved at the same time through the index.

[0104] Specifically, when each data is stored in the database, the final combined feature vector is obtained by calculating the original data, thereby constructing the index information of the original data; and the index information is updated in each data in the database.

[0105] The method of simultaneously retrieving the index-tagged data through the index includes: when querying data A, data A itself has multiple index information, and capturing the tag features pointed to by the index; when outputting data A, the captured tag features are simultaneously retrieved and output together.

[0106] It should be noted that in the data retrieval process, the data marked by the index can be retrieved at the same time through the index information, which can ensure that when a certain data is queried, not only the data itself but also all the tag features related to it can be obtained. The role of this design is that when querying data A, multiple index information of data A points to different tag features, which can capture the tag features pointed to by these indexes and output them together when outputting data A. This mechanism makes the query results more comprehensive and detailed, and helps to better understand the background and association of the data. For example, in the management of power plant operation data, when querying the data of a certain equipment, the response characteristics and variation of the equipment under different operating conditions can be obtained at the same time, thereby providing richer information support for fault diagnosis, operation optimization, etc.

[0107] Another thing to mention is that, for example, data a is the original data, and data c, data d, and data e are indexed. Then when we retrieve data c, we can retrieve data a at the same time according to the index. This design can ensure that when retrieving data for management, relevant data can be retrieved to the maximum extent, providing a data basis for management, prediction, and other behaviors.

[0108] On the other hand, this embodiment also provides a data marking system based on database intelligent management, which includes:

[0109] Acquisition unit, data acquisition, including obtaining raw data from multiple data sources.

[0110] The processing unit assigns data features to the original data and generates variation features according to the original data; performs feature reconstruction according to the variation situation and generates a labeling result of the original data.

[0111] The query unit stores the original data in the database and updates the corresponding index information; when retrieving data, the data marked by the index is retrieved at the same time through the index.

[0112] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0113] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0114] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0115] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0116] Example 2, below, is an embodiment of the present invention, which provides a data labeling method based on database intelligent management. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0117] Experiment 1: This experimental environment includes a server with the same configuration (CPU: Intel Xeon, RAM: 32GB, Storage: SSD), running the traditional database management method (using basic database storage, simple indexing and static rules for data retrieval and linkage retrieval) and the method of the present invention respectively. The data source simulates the operating data of the power plant, including temperature, pressure, flow and voltage, and the data volume is set to 1000, 5000, 10000, 50000 and 100000. A standard value is preset for each data type, and the deviation is calculated in combination with the current value, and the neural network is further used to predict the response of other data and equipment that can be caused.

[0118] Obtain raw data from the simulated power plant data source, and capture the data source information of the input data when entering the data into the database. The input data and its data source are regarded as raw data.

[0119] When each data is stored in the database, the final combined feature vector is obtained by calculating the original data, and the index information of the original data is constructed. The index information includes the data source, the original data, and the timestamp. The index information is updated in each data in the database to ensure the real-time and accuracy of the index.

[0120] When querying data A, the index information is used to retrieve the data marked by the index at the same time, and the marking features pointed to by the index are captured. Compared with the traditional method, the present invention achieves more efficient data retrieval and linkage retrieval through more complete and higher-quality index marking. When outputting data A, the captured marking features are retrieved and output together, and the retrieval time and the amount of data retrieved in linkage are recorded. The data record is shown in Table 1.

[0121] Table 1 Data comparison table

[0122]

[0123] From Table 1, it can be clearly seen that the present invention is significantly superior to the traditional method in many aspects: as the amount of data increases, the present invention is always superior to the traditional method in data storage time. For example, when storing 100,000 pieces of data, the traditional method takes 90 seconds, while the present invention only takes 60 seconds. This shows that the present invention significantly improves the efficiency of data storage by optimizing the data structure and storage strategy during data storage. The present invention is also significantly superior to the traditional method in index update time. For 100,000 pieces of data, the traditional method takes 50 seconds to update the index, while the present invention only takes 20 seconds. This shows that the present invention greatly reduces the time for index update through an automated index update mechanism, and improves the real-time performance and response speed of the system. The comparison of data retrieval time also shows the advantages of the present invention. For example, when retrieving 100,000 pieces of data, the traditional method takes 120 seconds, while the present invention only takes 70 seconds. This reflects that the present invention achieves more efficient data retrieval and significantly shortens the retrieval time through better index tags. The present invention also shows better performance in the amount of linked data retrieval. For example, in a scenario with 100,000 data items, the amount of data retrieved by the traditional method is 1,500, while the present invention can retrieve 1,600. This shows that the present invention has stronger performance in data linkage and relevance, and can capture relevant data more comprehensively.

[0124] Experiment 2: Query scenario setting: Temperature data of a key equipment (such as a boiler) in a power plant at a specific time point Abnormal deviation .

[0125] Use the built dynamic index to quickly locate all data items related to temperature data. Based on the abnormal deviation of temperature data, all related data and device status data are marked with index information. For example, temperature deviation Can cause stress Changes and flow changes.

[0126] Test: Through the method of the present invention, the query pressure Change or flow When the temperature changes, the relevant temperature data can be retrieved simultaneously. Together, this data provides a snapshot of the plant’s operating status at a specific point in time and reveals the impact of temperature anomalies on other parts of the system.

[0127] 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 preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A data marking method based on database intelligent management, characterized in that: include: Data collection, including obtaining raw data from multiple data sources; Assigning data features to the original data, and generating variation features based on the original data; According to the variation situation, feature reconstruction is performed to generate a labeling result of the original data; The original data is stored in the database and the corresponding index information is updated; when retrieving data, the data marked by the index is retrieved at the same time through the index; The data features include presetting standard values ​​for each type of data, combining current values, and calculating deviations: D ij =d ij -S ij Among them, d ij Indicates the jth data in the i-th data source, which is the current value; S ij Indicates the preset standard value of the jth data in the i-th data source; Δ ij Indicates the deviation of the jth data in the i-th data source; Assume that each data can cause the response of other data and other devices except itself; for the current data d ij The standard value and its deviation, using neural network to predict other data responses and other equipment responses that can be caused; Input Layer: X ij =[d ij ,S ij ,D ij ] Hidden layer calculation: H (l) =f(W (l) ·H (l-1) +b (l) ) Output layer calculation: P ij =W p ·H (L) +b p E ij =W e ·H (L) +b e Combined eigenvectors: F ij ={S ij ,D ij ,P ij ,E ij } Among them, X ij Represents the input of the neural network; H (l-1) represents the output of the l-1th hidden layer of the network; W (l) represents the weight matrix of the lth layer of the neural network; b (l) represents the bias vector of the lth layer of the neural network, W p and W e Respectively represent the weight matrices of parameter changes and equipment changes; b p and b e denotes the bias vectors of parameter changes and device changes respectively; f(...) denotes the activation function of ReLU; L denotes the total number of hidden layers; P ij Represents the current value of the jth data in the ith data source, which is the data set that can cause a response; E ij represents the current value of the jth data in the i-th data source, and the set of devices that can cause a response; F ij represents the feature vector; The feature reconstruction includes using a neural network to recalculate the combined feature vector to obtain a mutated combined feature vector F' ij , take the union of the combined feature vectors before and after the mutation to form the final combined feature vector: in, Represents the current value of the jth data in the i-th data source, the final set of data that can cause a response; P' ij Represents the current value of the jth data in the i-th data source, and the set of data mutations that can cause a response; Indicates the current value of the jth data in the i-th data source, and the final set of devices that can cause a response; E' ij Represents the current value of the jth data item in the i-th data source, and the set of device mutations that can cause a response; The labeling results include: And the final collection of equipment Mark each element in the data source, the original data and the timestamp; At the same time, index information containing the tag features is generated according to the tag, and the data source, original data and timestamp can be obtained through the index information.

2. The data marking method based on database intelligent management according to claim 1, characterized in that: The raw data includes, when the data information generated in the power plant is input into the database, the data source information of the input data is captured at the same time; The input data and its data source are regarded as raw data, and the input data are initially classified according to the data source; Original data format: D i ={(d i1 ,s1),(d i2 ,s2),...,(d in ,s n )} Among them, d in Indicates the nth data in the ith data source, s n Indicates the data source of the nth piece of data.

3. The data marking method based on database intelligent management as claimed in claim 2, characterized in that: The variation features include: for each original data d ij , based on the historical records of the same data, calculate the coefficient of variation of the data: Historical average: Standard Deviation: Coefficient of Variation: For the original data d ij To mutate: d’ ij =d ij [1+rand(0,CV ij )] Among them, d' ij Represents the original data after mutation, rand(0,CV ij ) means 0 to CV ij A random number between ij ) means d ij The maximum value of the same type of data, t represents the index of the record, and T represents the number of records.

4. The data marking method based on database intelligent management as claimed in claim 3, characterized in that: The updating of the corresponding index information includes, when each data is stored in the database, obtaining the final combined feature vector by calculating the original data, thereby constructing the index information of the original data; And in each data in the database, update the index information; The method of simultaneously retrieving the index-tagged data through the index includes: when querying data A, data A itself has multiple index information, and capturing the tag features pointed to by the index; when outputting data A, the captured tag features are simultaneously retrieved and output together.

5. A data marking system based on database intelligent management using the method according to any one of claims 1 to 4, characterized in that: Acquisition unit, data acquisition, including obtaining raw data from multiple data sources; A processing unit assigns data features to the original data, and generates variation features according to the original data; performs feature reconstruction according to the variation, and generates a labeling result of the original data; The query unit stores the original data in the database and updates the corresponding index information; when retrieving data, the data marked by the index is retrieved at the same time through the index.

6. A computer device comprising: Memory and processor; The memory stores a computer program, wherein the processor implements the steps of any one of the methods of claims 1 to 4 when executing the computer program.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images