Data quality enhancement method, system and equipment and storage medium
By identifying perceived weak areas and using RBM network and multi-assumption prediction residual reconstruction algorithm, the problem of unstable data quality in the group intelligence collaborative perception network is solved, and data quality improvement and data optimization of perceived weak areas are achieved, which enhances the Model reliability.
Patent Information
- Application Number
- CN202410884828.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-03
- Publication Date
- 2024-11-08
AI Technical Summary
The data quality in the group intelligence collaborative perception network is unstable, resulting in uneven data acquisition in the perceived weak area, affecting the reliability of the model.
Through the perception platform, the RBM network is used to perform grid data and feature mapping, combined with multi-assumption prediction residual reconstruction algorithm, the decoder is used to restore grid data at this time, establish data structure mapping relationship and data recovery method. Integrate user location information to optimize data quality.
The data quality of group intelligence perceptual network is improved, the inference results of data content in perceived weak areas are enhanced, and the balance of data acquisition and the reliability of the model are improved.
Smart Images

Figure CN118916610A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data enhancement technology, and in particular to a data quality enhancement method, system, device and storage medium. Background Art
[0002] The collective intelligence collaborative perception originates from the idea of "crowdsourcing". It uses "participatory" perception to complete tasks through conscious or unconscious collaboration among groups. The collective intelligence collaborative perception network provides tools for information interaction and calculation for collective intelligence collaborative perception. It combines "people and mobile devices" and adopts a human-machine hybrid perception method to achieve real-time perception and calculation throughout the city, so that large-scale and complex perception tasks can be completed in a short time. The collective intelligence collaborative perception network is applied to multiple fields, including environmental monitoring, intelligent transportation, public safety, social recommendation, urban planning, network optimization, medical health, etc. However, collective intelligence collaborative perception is heavily dependent on user mobility and data quality. User mobility determines the balance of data collection, and data quality determines the reliability of the collective intelligence collaborative perception model. Summary of the invention
[0003] The present invention provides a data quality enhancement method, system, device and storage medium, which identifies the weak perception area through the perception platform, and then realizes the mapping between grid data and features with the help of RBM, and finally realizes the grid data feature correction with the help of multi-hypothesis prediction residual reconstruction algorithm and restores the grid data at this moment with the RBM reverse decoder. A data structure mapping relationship between non-uniformly sampled data and a data recovery residual reconstruction method are established to realize data structure feature extraction, data mapping and data recovery of fused user location information, optimize the inference results of the data content of the weak perception area through data structure mapping, and enhance the data quality of the crowd intelligence perception network.
[0004] The present invention provides a data quality enhancement method, comprising:
[0005] Collect sample data from the perception platform, divide the sample data into several clusters according to distance using a clustering algorithm, define clusters with sparse sample points as weak perception areas, and divide them into several grids; put the sample data into the grids, and assign values to the grids according to the number of sample data in a grid;
[0006] Input the grid values into the RBM network, perform feature mapping on the sample data, obtain data features, and use spatiotemporal association to enhance data quality; in the time dimension, there is correlation between data in adjacent time periods;
[0007] The time correlation of the same grid sampling data is used to predict the data characteristics of the next moment, and the bidirectional estimation prediction residual is used to reconstruct the data characteristics of the current moment; the decoder is used to restore the data of the current moment to optimize the data quality of different grids.
[0008] Preferably, the sample information includes longitude and latitude information.
[0009] Preferably, when collecting sample data of the perception platform, the sample data is divided into several clusters according to distance through a clustering algorithm, a cluster with sparse sample points is defined as a weak perception area, and several grids are divided; the sample data is included in the grid, and the grid is assigned a value according to the number of sample data in a grid:
[0010] The method of dividing the sample data into several clusters according to distance by using a clustering algorithm includes:
[0011] Based on the longitude and latitude information, the k-means clustering algorithm is used to divide the samples into K clusters according to the distance between the sample data; specifically:
[0012]
[0013] The smaller the E value, the higher the similarity of sample data within the cluster; i represents the i-th cluster, u i represents the center point of the i-th cluster, and x represents the longitude and latitude information;
[0014] The method of defining a cluster with sparse sample points as a weak perception area and dividing the cluster into a plurality of grids; incorporating sample data into the grids, and assigning values to the grids according to the number of sample data in the grids includes:
[0015] A cluster with sparse sample points is defined as a weak perception area, and the location of the area is divided into a 10×10 grid, and the sample data is included in the grid according to the longitude and latitude information; if there are multiple sample data in a grid, the sample data is averaged and assigned a value to the grid; if the grid lacks sample data, the data value of the grid is 0.
[0016] Preferably, the step of inputting the grid values into the RBM network, performing feature mapping on the sample data, obtaining data features and enhancing data quality by utilizing spatiotemporal associations includes:
[0017] The RBM network is used to extract the features of the sampled data. The RBM has L layers. From bottom to top, the weight matrices of the layer nodes are: W0, W1, ..., W L-2 , the biases are: b1, b2, ..., b L ;
[0018] The input grid is assigned as:
[0019] A0=(a1,a2,...,a t )
[0020] Then the output of the jth neuron in the second layer is:
[0021]
[0022] Take all the outputs of the second layer of neurons as input and calculate the output of the next layer of neurons, and so on; then the output of the jth neuron in the lth layer is:
[0023]
[0024] where a i is the output of the i-th neuron in layer l-1, that is, p(h (l-1)i =1), there are L layers in total, then the output of the jth neuron in the last layer is:
[0025]
[0026] Among them, a i = p(h (l-1)i =1), then the feature of the Lth layer can be expressed as:
[0027] Z T =[p(h L1 =1),...,p(h Lj =1),...,p(h Lm =1)].
[0028] Preferably, the method of using the time correlation of the same grid sampling data to predict the data characteristics of the next moment, and using the bidirectional estimation prediction residual reconstruction to correct the data characteristics of the current moment; using the decoder to restore the data of the current moment to achieve the optimization of the data quality of different grids includes:
[0029] On the basis of obtaining data features, the time correlation of the same grid sampling data is used to realize multi-hypothesis prediction of the data features of the next moment, and the multi-hypothesis prediction residual reconstruction is realized by using bidirectional estimation to correct the data features of the moment; let the current moment be T, combined with the data features Z of the previous moment T-1 and the data feature Z at the next moment T+1 The current moment data features Perform multi-hypothesis prediction. The multi-hypothesis prediction characteristics at time T are:
[0030]
[0031] The error of the predicted value is further corrected by using two-way estimation based on the time series, specifically:
[0032]
[0033] e T Represents the prediction error at time T; Reconstruct the data features at time T by combining the data features obtained by RBM and the residual It is expressed as:
[0034]
[0035] Get the data features at time T Finally, the decoder is used to recover the data at time T and perform reverse transfer between the visible layer and hidden layer of the RBM to optimize the data quality of different grids.
[0036] Preferably, a data quality enhancement system comprises:
[0037] The data collection module is used to collect sample data from the perception platform, divide the sample data into several clusters according to distance through a clustering algorithm, define clusters with sparse sample points as weak perception areas, and divide them into several grids; put the sample data into the grid, and assign values to the grid according to the number of sample data in a grid;
[0038] The data mapping module is used to input the grid values into the RBM network, perform feature mapping on the sample data, obtain data features, and enhance data quality by using spatiotemporal associations; in the time dimension, data in adjacent time periods are correlated;
[0039] The data reconstruction module is used to use the time correlation of the same grid sampling data to predict the data characteristics of the next moment, and use the bidirectional estimation prediction residual reconstruction to correct the data characteristics of the current moment; the decoder is used to restore the data of the current moment to optimize the data quality of different grids.
[0040] Preferably, a device comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the data quality enhancement method when executing the computer program.
[0041] Preferably, a computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the data quality enhancement method.
[0042] Compared with the prior art, the present invention has the following beneficial effects:
[0043] The present invention discloses a data quality enhancement method, system, device and storage medium, which identifies the weak perception area through the perception platform, and then realizes the mapping between grid data and features with the help of RBM, and finally realizes the grid data feature correction with the help of multi-hypothesis prediction residual reconstruction algorithm and restores the grid data at this moment with the RBM reverse decoder. A data structure mapping relationship and a data recovery residual reconstruction method between non-uniformly sampled data are established to realize data structure feature extraction, data mapping and data recovery of fused user location information, optimize the inference results of the data content of the weak perception area through data structure mapping, and enhance the data quality of the crowd intelligence perception network. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] Figure 1 It is a flow chart of a data quality enhancement method provided by an embodiment of the present invention;
[0045] Figure 2 is a schematic diagram of the RBM network structure provided by an embodiment of the present invention;
[0046] Figure 3 is a schematic diagram of a weak perception area provided by an embodiment of the present invention;
[0047] Figure 4 is a schematic diagram of a data quality enhancement system module provided by an embodiment of the present invention;
[0048] Figure 5 It is a schematic diagram of a data quality management framework provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0049] 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.
[0050] See also Figure 1 , a data quality enhancement method, comprising:
[0051] S1: Collect sample data from the perception platform, divide the sample data into several clusters according to distance using a clustering algorithm, define clusters with sparse sample points as weak perception areas, and divide them into several grids; put the sample data into the grids, and assign values to the grids according to the number of sample data in a grid;
[0052] S2: Input the grid values into the RBM network, perform feature mapping on the sample data, obtain data features, and use spatiotemporal association to enhance data quality; in the time dimension, there is correlation between data in adjacent time periods;
[0053] S3: Use the time correlation of the same grid sampling data to predict the data characteristics of the next moment, and use the bidirectional estimation prediction residual reconstruction to correct the data characteristics of the current moment; use the decoder to restore the data of the current moment to optimize the data quality of different grids.
[0054] Preferably, the sample information includes longitude and latitude information.
[0055] Preferably, in step S1, after collecting sample data of the perception platform, the sample data is divided into several clusters according to distance by a clustering algorithm, a cluster with sparse sample points is defined as a perception weak area, and several grids are divided; the sample data is included in the grid, and the grid is assigned a value according to the number of sample data in a grid:
[0056] The method of dividing the sample data into several clusters according to distance by using a clustering algorithm includes:
[0057] Based on the longitude and latitude information, the k-means clustering algorithm is used to divide the samples into K clusters according to the distance between the sample data; specifically:
[0058]
[0059] The smaller the E value, the higher the similarity of sample data within the cluster; i represents the i-th cluster, u i represents the center point of the i-th cluster, and x represents the longitude and latitude information;
[0060] In the above scheme, the core idea is to make the points of the same cluster as closely connected as possible, and make the distance between clusters as large as possible, so as to minimize the square error of the divided cluster C; the clustering algorithm formula describes the closeness of the samples in the cluster around the cluster mean vector. The smaller the E value, the higher the similarity of the samples in the cluster.
[0061] The method of defining a cluster with few sample points as a weak perception area, dividing the weak perception area into a plurality of grids, incorporating sample data into the grids, and assigning values to the grids according to the number of sample data in a grid includes:
[0062] A cluster with sparse sample points is defined as a weak perception area, and the location of the area is divided into a 10×10 grid, and the sample data is included in the grid according to the longitude and latitude information; if there are multiple sample data in a grid, the sample data is averaged and assigned a value to the grid; if the grid lacks sample data, the data value of the grid is 0.
[0063] Preferably, in step S2, the step of inputting the grid values into the RBM network, performing feature mapping on the sample data, obtaining data features and enhancing data quality by utilizing spatiotemporal association includes:
[0064] The RBM network is used to extract the features of the sampled data. The RBM has L layers. From bottom to top, the weight matrices of the layer nodes are: W0, W1, ..., W L-2 , the biases are: b1, b2, ..., b L ;
[0065] The input grid is assigned as:
[0066] A0=(a1,a2,...,a t )
[0067] Then the output of the jth neuron in the second layer is:
[0068]
[0069] Take all the outputs of the second layer of neurons as input and calculate the output of the next layer of neurons, and so on; then the output of the jth neuron in the lth layer is:
[0070]
[0071] where a i is the output of the i-th neuron in layer l-1, that is, p(h (l-1)i =1), there are L layers in total, then the output of the jth neuron in the last layer is:
[0072]
[0073] Among them, a i = p(h (l-1)i =1), then the feature of the Lth layer can be expressed as:
[0074] Z T =[p(h L1 =1),...,p(h Lj =1),...,p(h Lm =1)].
[0075] In the above scheme, see Figure 2 , the conventional RBM network consists of 2 layers, the first layer is the visible layer, and the second layer is the hidden layer. The visible layer consists of t visible nodes, the hidden layer consists of m hidden nodes, and w is the weight matrix of visible nodes and hidden nodes of order t×m;
[0076] Assume the bias of the visible node of the visual layer is: b = (b1, b2, ..., b t );
[0077] Assume the bias of the hidden node of the hidden layer is: c = (c1, c2, ..., c m );
[0078] The characteristic value of the hidden node obtained by assigning the visible node input grid is:
[0079]
[0080] In order to obtain high-dimensional semantics of the sampled data, the present invention adopts multi-layer RBM to realize feature extraction.
[0081] Preferably, in step S3, the method of using the time correlation of the same grid sampling data to predict the data characteristics at the next moment, and using the bidirectional estimation prediction residual reconstruction to correct the data characteristics at the current moment; using the decoder to restore the data at the current moment to optimize the data quality of different grids includes:
[0082] On the basis of obtaining data features, the time correlation of the same grid sampling data is used to realize multi-hypothesis prediction of the data features of the next moment, and the multi-hypothesis prediction residual reconstruction is realized by using bidirectional estimation to correct the data features of the moment; let the current moment be T, combined with the data features Z of the previous moment T-1 and the data feature Z at the next moment T+1 The current moment data features Perform multi-hypothesis prediction. The multi-hypothesis prediction characteristics at time T are:
[0083]
[0084] The error of the predicted value is further corrected by using two-way estimation based on the time series, specifically:
[0085]
[0086] e T Represents the prediction error at time T; Reconstruct the data features at time T by combining the data features obtained by RBM and the residual It is expressed as:
[0087]
[0088] Get the data features at time T Finally, the decoder is used to recover the data at time T and perform reverse transfer between the visible layer and hidden layer of the RBM to optimize the data quality of different grids.
[0089] In one embodiment, a swarm intelligence collaborative perception network is used to perceive the communication quality of a mobile communication network, and a mobile user complaint database is constructed by extracting features of signaling information of dissatisfied users, thereby improving the accuracy of end-to-end mobile network service quality assessment based on the swarm intelligence collaborative perception network.
[0090] This paper selects 10,000 users as sensing nodes and obtains the signaling data of the user's time and space distribution through the user's daily movement. The user will directly call the complaint phone during the movement. The complaint agency will obtain the signaling information of the time and location based on the time and location of the user's call, thereby realizing the association between user complaint data and mobile network quality indicators. Table 1 shows the mobile network quality indicators obtained based on user complaint data.
[0091] Table 1 Mobile network quality indicators
[0092]
[0093]
[0094] See also Figure 3 On the basis of obtaining the signaling of the complaining users, considering the uniformity and dynamics of mobile users, this paper starts with the three steps of data collection, data mapping and data reconstruction, and uses the k-means clustering algorithm to identify the weak perception area; then the location of the area is divided into 10×10 grids, and the samples are included in the grid according to the longitude and latitude information.
[0095] This paper verifies the rationality of the algorithm in this paper. This paper reconstructs the mobile network quality index data of user complaints by combining the spatial coverage analysis model, sparse reconstruction method and the method in this paper, and uses the BP neural network to associate with user complaints to build a user complaint prediction model to realize user complaint prediction.
[0096] Preferably, see Figure 4 , a data quality enhancement system, comprising
[0097] The data collection module is used to collect sample data from the perception platform, divide the sample data into several clusters according to distance through a clustering algorithm, define clusters with sparse sample points as weak perception areas, and divide them into several grids; put the sample data into the grid, and assign values to the grid according to the number of sample data in a grid;
[0098] The data mapping module is used to input the grid values into the RBM network, perform feature mapping on the sample data, obtain data features, and enhance data quality by using spatiotemporal associations; in the time dimension, data in adjacent time periods are correlated;
[0099] The data reconstruction module is used to use the time correlation of the same grid sampling data to predict the data characteristics of the next moment, and use the bidirectional estimation prediction residual reconstruction to correct the data characteristics of the current moment; the decoder is used to restore the data of the current moment to optimize the data quality of different grids.
[0100] In one embodiment, see Figure 5 , data quality management is divided into three modules. The first module is the collection end, which collects sample data from the perception platform and divides the sample data into several clusters according to the distance through the clustering algorithm. The clusters with sparse sample points are defined as the perception weak area and divided into several grids; the sample data is included in the grid, and the grid is assigned according to the number of sample data in a grid. The second module is the data mapping end, which is responsible for feature mapping of the data, obtaining data features and enhancing data quality by using spatiotemporal association. In the time dimension, there is correlation between data in adjacent time periods; in the spatial dimension, there is correlation between data between adjacent grids. The missing data features of some grids at a certain moment can be preliminarily reconstructed through the data features of the previous and next moments. The third module is the reconstruction end, which is responsible for complementary data enhancement, using the bidirectional estimation method to achieve residual reconstruction of data features at a certain moment, forming a re-correction of the gridded data features of the perception weak area, and then using the decoder to restore the data at a certain moment to achieve data enhancement in the weak area.
[0101] The present invention also provides a device, which is a structural schematic diagram of a preferred embodiment of the device provided by the present invention, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, it implements a data quality enhancement method described in any of the above embodiments.
[0102] It should be noted that, in some specific embodiments, the device may further include multiple memories and / or multiple processors, and their specific number and connection method may be set and adaptively adjusted according to actual needs.
[0103] The present invention also provides a computer-readable storage medium, specifically including a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a data quality enhancement method as described in any of the above embodiments.
[0104] It should be noted that the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be further noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.
[0105] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.
Claims
1. A data quality enhancement method, characterized in that: include: Collect sample data from the perception platform, divide the sample data into several clusters according to distance using a clustering algorithm, define clusters with sparse sample points as weak perception areas, and divide them into several grids; put the sample data into the grids, and assign values to the grids according to the number of sample data in a grid; Input the grid values into the RBM network, perform feature mapping on the sample data, obtain data features, and use spatiotemporal association to enhance data quality; in the time dimension, there is correlation between data in adjacent time periods; The time correlation of the same grid sampling data is used to predict the data characteristics of the next moment, and the bidirectional estimation prediction residual is used to reconstruct the data characteristics of the current moment; the decoder is used to restore the data of the current moment to optimize the data quality of different grids.
2. A data quality enhancement method according to claim 1, characterized in that: When collecting sample data from the perception platform, the sample data is divided into several clusters according to distance through a clustering algorithm. Clusters with sparse sample points are defined as weak perception areas and divided into several grids. The sample data is included in the grids, and the grids are assigned values according to the number of sample data in a grid: The method of dividing the sample data into several clusters according to distance by using a clustering algorithm includes: Based on the sample data, the k-means clustering algorithm is used to divide the samples into K clusters according to the distance between the sample data; specifically: The smaller the E value, the higher the similarity of sample data within the cluster; i represents the i-th cluster, u i represents the center point of the i-th cluster, and x represents the longitude and latitude information; The method of defining a cluster with sparse sample points as a weak perception area and dividing the cluster into a plurality of grids; incorporating sample data into the grids, and assigning values to the grids according to the number of sample data in the grids includes: A cluster with sparse sample points is defined as a weak perception area, and the location of the area is divided into a 10×10 grid, and the sample data is included in the grid; if there are multiple sample data in a grid, the sample data is averaged and assigned a value to the grid; if the grid lacks sample data, the data value of the grid is 0.
3. A data quality enhancement method according to claim 2, characterized in that: The sample data includes longitude and latitude information.
4. A data quality enhancement method according to claim 2, characterized in that: The step of inputting the grid values into the RBM network, performing feature mapping on the sample data, obtaining data features and enhancing data quality by using spatiotemporal association includes: The RBM network is used to extract the features of the sampled data. The RBM has L layers. From bottom to top, the weight matrices of the layer nodes are: W0, W1, ..., W L-2 , the biases are: b1, b2, ..., b L ; The input grid is assigned as: <h2 style=";text-align:left;direction:ltr">A0 = (a1, a2,..., a)<h2 style=";text-align:left;direction:ltr"> t <h2 style=";text-align:left;direction:ltr"> ) Then the output of the jth neuron in the second layer is: Take all the outputs of the second layer of neurons as input and calculate the output of the next layer of neurons, and so on; then the output of the jth neuron in the lth layer is: where a i is the output of the i-th neuron in layer l-1, that is, p(h ( l -1)i =1), there are L layers in total, then the output of the jth neuron in the last layer is: Among them, a i = p(h (l-1)i =1), then the feature of the Lth layer can be expressed as: Z T =[p(h L1 =1),...,p(h Lj =1),...,p(h Lm =1)]。 5. A data quality enhancement method according to claim 4, characterized in that: The method uses the time correlation of the same grid sampling data to predict the data characteristics of the next moment, and uses the bidirectional estimation prediction residual reconstruction to correct the data characteristics of the current moment; The decoder is used to restore the data at this moment to optimize the data quality of different grids, including: On the basis of obtaining data features, the time correlation of the same grid sampling data is used to realize multi-hypothesis prediction of the data features of the next moment, and the multi-hypothesis prediction residual reconstruction is realized by using bidirectional estimation to correct the data features of the moment; let the current moment be T, combined with the data features Z of the previous moment T-1 and the data feature Z at the next moment T+1 The current moment data features Perform multi-hypothesis prediction. The multi-hypothesis prediction characteristics at time T are: The error of the predicted value is further corrected by using two-way estimation based on the time series, specifically: e T Represents the prediction error at time T; Reconstruct the data features at time T by combining the data features obtained by RBM and the residual It is expressed as: Get the data features at time T Finally, the decoder is used to recover the data at time T and perform reverse transfer between the visible layer and hidden layer of the RBM to optimize the data quality of different grids.
6. A data quality enhancement system, characterized in that: include The data collection module is used to collect sample data from the perception platform, divide the sample data into several clusters according to distance through a clustering algorithm, define clusters with sparse sample points as weak perception areas, and divide them into several grids; put the sample data into the grid, and assign values to the grid according to the number of sample data in a grid; The data mapping module is used to input the grid values into the RBM network, perform feature mapping on the sample data, obtain data features, and enhance data quality by using spatiotemporal associations; in the time dimension, data in adjacent time periods are correlated; The data reconstruction module is used to use the time correlation of the same grid sampling data to predict the data characteristics of the next moment, and use the bidirectional estimation prediction residual reconstruction to correct the data characteristics of the current moment; the decoder is used to restore the data of the current moment to optimize the data quality of different grids.
7. A device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, a data quality enhancement method according to any one of claims 1 to 5 is implemented.
8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute a data quality enhancement method as claimed in any one of claims 1 to 5.
