Mass data set quality evaluation method and device, electronic equipment and storage medium

Abstract

The invention provides a mass data set quality evaluation method and device, electronic equipment and a storage medium. Comprising the steps of determining a quality evaluation index of a target data set based on the type of the target data set; determining a quality evaluation matrix based on index data, corresponding to each quality evaluation index, of a data sample in the target data set, and preprocessing the quality evaluation matrix to obtain a standard quality evaluation matrix; determining the entropy weight of each quality evaluation index based on the standard quality evaluation matrix and an entropy weight method, and weighting the standard quality evaluation matrix based on the entropy weight to obtain a weighted standard matrix; and based on the weighting standard matrix and an ideal solution, determining the relative proximity of the data samples in the target data set, and based on the relative proximity, determining a quality evaluation result of the target data set. By adopting the technical scheme, the problems that an evaluation mode in the prior art cannot evaluate the quality performance of the whole data set and the quality evaluation effect of the data set is poor are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of data set quality assessment, and in particular to a method, device, electronic device and storage medium for assessing the quality of a massive data set. Background Art

[0002] With the rapid development of remote sensing technology, remote sensing data is growing rapidly, and data quality issues are becoming increasingly prominent. During data collection, transmission, and processing, quality issues such as no labeling, noise, and incompleteness may occur, leading to serious errors in data-based decision-making. Data quality assessment has become an indispensable link.

[0003] Existing data quality assessment technologies mostly focus on structured data and the assessment of individual unstructured data (such as a single image). There is a lack of assessment methods for the overall quality of a data set (such as 10,000 texts, 1 million images), resulting in poor overall quality assessment of the data set. Summary of the invention

[0004] The present invention provides a mass data set quality assessment method, device, electronic device and storage medium, which are used to solve the problem that the quality performance of the entire data set cannot be assessed using the assessment method in the prior art and the assessment effect is poor.

[0005] In a first aspect, the present invention provides a quality assessment method for a massive data set, comprising: determining a quality assessment indicator of a target data set based on a type of the target data set, the type of the target data set including data sets of the same type and data sets of different types; determining a quality assessment matrix based on indicator data corresponding to each quality assessment indicator of data samples in the target data set, and preprocessing the quality assessment matrix to obtain a standard quality assessment matrix; determining the entropy weight of each quality assessment indicator based on the standard quality assessment matrix and an entropy weight method, and weighting the standard quality assessment matrix based on the entropy weight to obtain a weighted standard matrix; determining the relative proximity of data samples in the target data set based on the weighted standard matrix and an ideal solution, and determining a quality assessment result of the target data set based on the relative proximity.

[0006] Optionally, the quality assessment index of the target dataset is determined based on the type of the target dataset, including: when the target datasets are datasets of the same type, using the overall quality index and / or intrinsic quality index of the datasets of the same type as the quality assessment index of the target dataset; when the target datasets are datasets of different types, using the overall quality index of datasets of different types as the quality assessment index of the target dataset.

[0007] Optionally, overall quality indicators include: data completeness, data content diversity, data timeliness, retrieval efficiency and user satisfaction.

[0008] Optionally, the target data set is a data set of remote sensing data, and the remote sensing data includes at least one of digital orthophoto data, digital elevation model data, vector terrain framework data, navigation electronic map data and real-scene image data.

[0009] Optionally, the same type of data sets are data sets including any one of digital orthophoto data, digital elevation model data, vector terrain framework data, navigation electronic map data and real-life image data, and the different types of data sets are data sets including at least two of digital orthophoto data, digital elevation model data, vector terrain framework data, navigation electronic map data and real-life image data.

[0010] Optionally, the method of preprocessing the quality assessment matrix includes at least one of the following: unifying the indicator data with different indicator directions in the quality assessment matrix; dimensionlessly processing the indicator data with different dimensions in the quality assessment matrix by standard deviation method, extreme value difference method or efficacy coefficient method; quantizing the qualitative indicator data in the quality assessment matrix by linear quantization, partial large quantization or membership function.

[0011] In a second aspect, the present invention provides a quality assessment device for a massive data set, comprising: a first determination module, used to determine the quality assessment index of a target data set based on the type of the target data set, the type of the target data set including data sets of the same type and data sets of different types; a first processing module, used to determine a quality assessment matrix based on the index data of each quality assessment index corresponding to the data samples in the target data set, and pre-process the quality assessment matrix to obtain a standard quality assessment matrix; a second processing module, used to determine the entropy weight of each quality assessment index based on the standard quality assessment matrix and an entropy weight method, and weight the standard quality assessment matrix based on the entropy weight to obtain a weighted standard matrix; a second determination module, used to determine the relative proximity of the data samples in the target data set based on the weighted standard matrix and an ideal solution, and determine the quality assessment result of the target data set based on the relative proximity.

[0012] In a third aspect, the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the quality assessment method for a massive data set as described above is implemented.

[0013] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned methods for assessing the quality of a massive data set.

[0014] In a fifth aspect, the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned methods for assessing the quality of a massive data set.

[0015] The quality assessment method, device, electronic device and storage medium of the massive data set provided by the present invention determine the quality assessment index corresponding to the target data set by the type of the target data set, so that each target data set has its own corresponding quality assessment index, and then determine the quality assessment matrix based on the index data corresponding to each quality assessment index of the data samples in the target data set, and pre-process the quality assessment matrix to obtain a standard quality assessment matrix, eliminating the directional differences and scale differences between the index data. Further, the entropy weight method is used to determine the entropy weight of each quality assessment index in the standard quality assessment matrix to ensure the objectivity of the weight distribution, and the standard quality assessment matrix is ​​weighted based on the entropy weight to obtain a weighted standard matrix. Then, an ideal solution is used for the weighted standard matrix to determine the relative proximity of the data samples in the target data set, and the quality assessment result of the target data set is determined based on the relative proximity, so that the quality assessment result can better reflect the overall quality of the target data set. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a flow chart of the quality assessment method of a massive data set provided by the present invention.

[0018] Figure 2 It is a quality assessment index system diagram of the massive data set provided by the present invention.

[0019] Figure 3 It is a flow chart of a quality assessment method for a massive data set based on an entropy weight method and an ideal solution provided by the present invention.

[0020] Figure 4 It is a schematic diagram of the structure of the quality assessment device for massive data sets provided by the present invention.

[0021] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are 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.

[0023] Figure 1 FIG. 1 is a flow chart of the quality assessment method of a massive data set provided by the present invention. Figure 1 As shown, the method includes the following steps S101-S104.

[0024] Step S101: determining a quality assessment index of a target data set based on the type of the target data set, where the type of the target data set includes data sets of the same type and data sets of different types.

[0025] It should be noted that the judgment criterion for the same type of data sets and different types of data sets is whether the data in the target data set are of the same type.

[0026] As an optional embodiment, if the target dataset 1 only includes digital orthophoto data, the target datasets are datasets of the same type. If the target dataset 2 includes digital orthophoto data and digital elevation model data, the target datasets are datasets of different types.

[0027] It should be noted that the quality assessment indicators of the target data set include overall quality indicators and intrinsic quality indicators.

[0028] In an exemplary embodiment, the target data set is a data set of remote sensing data, and the remote sensing data includes at least one of digital orthophoto data, digital elevation model data, vector terrain framework data, navigation electronic map data, and real-scene image data.

[0029] In an exemplary embodiment, the same type of data sets are data sets that include any one of digital orthophoto data, digital elevation model data, vector terrain framework data, navigation electronic map data and real-life image data, and the different types of data sets are data sets that include at least two of digital orthophoto data, digital elevation model data, vector terrain framework data, navigation electronic map data and real-life image data.

[0030] As an optional embodiment, when the target data set only includes digital orthophoto data, the target data set is a data set of the same type. When the target data set includes digital elevation model data and vector terrain framework data, the target data set is a data set of different types. That is to say, when the remote sensing data in the target data set are data of the same type, the target data set is a data set of the same type, and when the remote sensing data in the target data set are data of different types, the target data set is a data set of different types. It should be further explained that the data set of the same type means that the data in the data set are data of the same type, and the data of the same type have the same quality indicators, but the quality indicators are heterogeneous. Similarly, the data sets of different types mean that the data in the data set are not data of the same type, and each type of data has its own quality indicators, and the evaluation criteria and dimensions of each quality indicator are different, that is, the quality indicators are heterogeneous. There is at least one type of remote sensing data in the target data set, and the number of remote sensing data in the target data set is not further limited by the present invention.

[0031] In an exemplary embodiment, determining a quality assessment indicator of a target data set based on the type of the target data set includes: When the target data set is a data set of the same type, the overall quality index and / or the intrinsic quality index of the data set of the same type is used as the quality assessment index of the target data set.

[0032] When the target datasets are datasets of different types, the overall quality indicators of the datasets of different types are used as the quality evaluation indicators of the target dataset.

[0033] As an optional embodiment, the target data set composed of different remote sensing data has different intrinsic quality indicators, and these indicators are heterogeneous. The heterogeneity can be reflected in the dimensions and evaluation methods.

[0034] In an exemplary embodiment, the overall quality indicators include: data completeness, data content diversity, data timeliness, retrieval efficiency, and user satisfaction.

[0035] As an optional embodiment, the various indicators of the overall quality indicator are also heterogeneous. For example, the evaluation standard of user satisfaction is satisfied or dissatisfied, and the evaluation standard of retrieval efficiency is a specific value.

[0036] It should be noted that data completeness refers to whether indicators such as the data volume and data coverage of the data set meet the indicator requirements.

[0037] As an optional embodiment, in the field of surveying and mapping, for orthophoto data, it is necessary to determine whether the orthophoto image with a resolution of 0.5 meters covers the global land area. For vector terrain framework data, it is necessary to determine whether the 1:10,000 vector terrain framework covers the target area. These indicators are all the contents to be checked for data integrity. The specific results can be calculated by computer software using indicators such as data volume and data coverage percentage, and the present invention does not make further limitations on this.

[0038] It should be noted that data content diversity refers to the relationship between various data units within a data set. The purpose is to prevent data providers from copying a small amount of data and using false data to meet the requirements of data indicators.

[0039] As an optional embodiment, the diversity of data content is measured by the similarity between two data units. For large-scale data sets, especially for unstructured data, it is unrealistic to test the similarity between data units one by one, and its time complexity and space complexity are unacceptable. Therefore, the locality-sensitive hashing (LSH), a sampling-based approximate nearest neighbor fast search technology for massive high-dimensional data, can be used to calculate the similarity between data units, thereby determining the quality index of data diversity. The diversity of data content is quantified by the average pairwise distance between the feature vectors of data points in the data set.

[0040] It should be noted that the data timeliness index reflects the accessibility of a data set and can be measured by keyword-based retrieval efficiency, fuzzy retrieval efficiency, and combined retrieval efficiency. The retrieval efficiency of a data set can be measured by the average return time and the average amount of returned data (such as data entries).

[0041] As an optional embodiment, user satisfaction is a qualitative indicator for measuring the quality of a data set. For example, if user satisfaction is {very dissatisfied, not very satisfied, relatively satisfied, satisfied, very satisfied}, these five levels can be respectively corresponded to 0.2, 0.4, 0.6, 0.8, and 1, which is a simple linear quantification. In actual application, user satisfaction is a specific quantitative value obtained by software statistics after obtaining user feedback on the data set. The quantification method and quantification scale are not further limited in the present invention.

[0042] As an optional embodiment, the intrinsic quality index is a quality index of the data itself other than the overall quality index, and the intrinsic quality indexes of the data are also heterogeneous. That is, the intrinsic quality index is a quality index of the data itself other than data completeness, data content diversity, data timeliness, retrieval efficiency and user satisfaction.

[0043] As an optional embodiment, the intrinsic quality index of the target data set composed of digital orthophoto data includes spatial reference system, position accuracy, time accuracy, image / grid quality and attachment quality. The index value of the intrinsic quality index of the digital orthophoto data can be determined by the inspection items and inspection contents. The intrinsic quality index and inspection items of the digital orthophoto data are shown in Table 1 below.

[0044] Table 1 - Intrinsic quality indicators and inspection items of digital orthophoto data

[0045] It should be noted that the target data set composed of digital orthophoto data already includes indicator data corresponding to the intrinsic quality indicators in the above table. The indicator data corresponding to the intrinsic quality indicators are obtained by checking the inspection contents corresponding to the inspection items in the above table. The inspection method and the method of obtaining the indicator data corresponding to the intrinsic quality indicators are not further limited in the embodiments of the present invention.

[0046] As an optional embodiment, the intrinsic quality index of the target data set composed of digital elevation model data includes spatial reference system, position accuracy, logical consistency, time accuracy, image / grid quality and attachment quality. The index value of the intrinsic quality index of the digital elevation model data can be determined by the inspection items and inspection contents. The intrinsic quality index and inspection items of the digital elevation model data are shown in Table 2 below.

[0047] Table 2 - Intrinsic quality indicators and check items of digital elevation model data

[0048] It should be noted that the target data set composed of digital elevation model data already includes indicator data corresponding to the intrinsic quality indicators in the above table. The indicator data corresponding to the intrinsic quality indicators are obtained by checking the inspection contents corresponding to the inspection items in the above table. The verification method and the method of obtaining the indicator data corresponding to the intrinsic quality indicators are not further limited in the embodiments of the present invention.

[0049] As an optional embodiment, the intrinsic quality index of the target data set composed of vector terrain framework data includes spatial reference system, position accuracy, attribute accuracy, integrity, logical consistency, time accuracy, representation quality and attachment quality. The index value of the intrinsic quality index of the vector terrain framework data can be determined by the inspection items and inspection contents. The intrinsic quality index and inspection items of the vector terrain framework data are shown in Table 3 below.

[0050] Table 3 - Intrinsic quality indicators and check items of vector terrain framework data

[0051] It should be noted that the target data set composed of the vector terrain framework data already includes the indicator data corresponding to the intrinsic quality indicators in the above table. The indicator data corresponding to the intrinsic quality indicators are obtained by checking the inspection contents corresponding to the inspection items in the above table. The inspection method and the method of obtaining the indicator data corresponding to the intrinsic quality indicators are not further limited in the embodiments of the present invention.

[0052] As an optional embodiment, the intrinsic quality index of the target data set composed of navigation electronic map data includes overall completeness, coordinate accuracy, normalization, data edge connection and manual interaction. The index data of the intrinsic quality index of the navigation electronic map data can be determined by inspection items and inspection contents. The intrinsic quality index and inspection items of the navigation electronic map data are shown in Table 4 below.

[0053] Table 4 - Intrinsic quality indicators and inspection items of navigation electronic map data

[0054] It should be noted that the target data set composed of the navigation electronic map data already includes the indicator data corresponding to the intrinsic quality indicators in the above table. The indicator values ​​corresponding to the intrinsic quality indicators are obtained by checking the inspection contents corresponding to the inspection items in the above table. The inspection method and the method of obtaining the indicator data corresponding to the intrinsic quality indicators are not further limited in the embodiments of the present invention.

[0055] As an optional embodiment, the intrinsic quality indicators of the target data set composed of real-scene image data include data integrity, data relevance, consistency, manual visualization inspection, intersection turning, consistency between real-scene image and road network, and image quality. The indicator data of the intrinsic quality indicators of the real-scene image data can be determined by inspection items and inspection contents. The intrinsic quality indicators and inspection items of the real-scene image data are shown in Table 5 below.

[0056] Table 5 - Intrinsic quality indicators and inspection items of real-life image data

[0057] It should be noted that the target data set composed of real-scene image data already includes indicator data corresponding to the intrinsic quality indicators in the above table. The indicator data corresponding to the intrinsic quality indicators are obtained by checking the inspection contents corresponding to the inspection items in the above table. The inspection method and the method of obtaining the indicator data corresponding to the intrinsic quality indicators are not further limited in the embodiments of the present invention.

[0058] It should be further explained that no matter which type of remote sensing data the target data set is composed of, it contains an overall quality index, and the intrinsic quality indicators corresponding to different remote sensing data are different. Therefore, for the same type of data sets, the overall quality index can be used as the quality evaluation index of the target data set, the intrinsic quality index can be used as the quality evaluation index of the target data set, and the intrinsic quality index and the overall quality index can be used together as the quality evaluation index of the target data set. The present invention does not further limit which quality index is specifically selected as the quality evaluation index of the target data set, and it can be flexibly selected according to actual needs. For different types of data sets, only the common overall quality index can be used as the quality evaluation index of the target data set.

[0059] Step S102: determining a quality assessment matrix based on the indicator data corresponding to each quality assessment indicator of the data samples in the target data set, and preprocessing the quality assessment matrix to obtain a standard quality assessment matrix.

[0060] As an optional embodiment, when determining the quality assessment matrix based on the index data corresponding to each quality assessment index of the data samples in the target data set, all data samples in the target data set can be selected to construct the quality assessment matrix, or some data samples can be selected to construct the quality assessment matrix. The present invention does not further limit the number of data samples, and can be flexibly selected according to the processing power of the computer and actual needs. At the same time, the present invention does not further limit the method for obtaining some data samples described herein, for example, a certain number of data samples are selected from the data samples in the target data set by random sampling to determine the quality assessment matrix.

[0061] As an optional embodiment, the preprocessing includes consistency processing, standardization processing, etc.

[0062] As an optional embodiment, assuming that the target data set is a data set of the same type, there are data samples, a total of quality assessment indicators, For the The data sample The indicator data of the quality assessment indicators ( ). Then the quality assessment matrix ( ) can be expressed as: It should be noted that each row in the quality assessment matrix corresponds to a data sample, and each column corresponds to a quality assessment indicator. It is the specific indicator data of the quality assessment indicator corresponding to a data sample.

[0063] It should be noted that the indicator data obtained from the target data set cannot be used directly for quality assessment, because there may be situations where the indicator data has different directions, different dimensions, or is a qualitative indicator without a specific value. Therefore, it is necessary to preprocess the indicator data in the quality assessment matrix.

[0064] In an exemplary embodiment, the quality assessment matrix is ​​preprocessed in at least one of the following ways: The indicator data with different indicator directions in the quality assessment matrix are consistent.

[0065] It should be noted that the directions of the indicators include positive indicators (the larger the value, the better), negative indicators (the smaller the value, the better), intermediate indicators (smaller values ​​are best) and interval indicators (values ​​within a certain range are best).

[0066] As an optional embodiment, for positive indicators, the larger the indicator, the better, so no processing is required. For negative indicators, the smaller the indicator, the better, so it can be converted into a positive indicator through an inverse transformation or a minimax transformation.

[0067] As an optional embodiment, for a centered indicator, the data is best near a middle value. is a set of centered indicator sequences with the middle value being , then the forward formula is: Represents an indicator data, Represents the indicator data after positive conversion.

[0068] As an optional embodiment, for interval indicators, the data is best within a certain interval. is a set of interval indicator sequences, and the best interval is , then the forward formula is: in, Indicates The data sample The indicator data of the quality assessment indicators, After positive The data sample The indicator data of the quality assessment indicators.

[0069] The indicator data with different dimensions in the quality assessment matrix are dimensionless processed through the standard deviation method, extreme value difference method or efficacy coefficient method.

[0070] It should be noted that if the indicator data needs to be dimensionless, it needs to be consistent first.

[0071] As an optional embodiment, the formula for dimensionless processing of the indicator data by the standard deviation is as follows: in, Represents a data sample, represents the quality assessment index, Indicates The mean of the indicator data corresponding to the quality assessment indicator, Indicates The standard deviation of the indicator data corresponding to each quality assessment indicator, and the mean and value of the standardized data are 0 and 1 respectively.

[0072] As an optional embodiment, the indicator data may also be linearly transformed (translated + scaled) to the interval [0, 1] by a normalization method. The normalization method is a commonly used mathematical method and will not be described in detail here.

[0073] The qualitative indicator data in the quality assessment matrix are quantified through linear quantization, partial large-scale quantization or membership function.

[0074] It should be noted that in practical applications, many indicators only have qualitative indicators but no quantitative indicators, such as user satisfaction. User satisfaction can be {very dissatisfied, not very satisfied, relatively satisfied, satisfied, very satisfied}. If qualitative indicators are used for quality assessment, the results of quality assessment will be inaccurate and cannot be quantitatively calculated. Therefore, it is necessary to quantify qualitative indicators.

[0075] As an optional embodiment, the qualitative indicator data may be linearly quantified. Taking user satisfaction as an example, {very dissatisfied, not very satisfied, relatively satisfied, satisfied, very satisfied} may correspond to 0.2, 0.4, 0.6, 0.8, and 1 respectively.

[0076] As an optional embodiment, a large-scale quantization may be performed on indicators such as {very bad, poor, average, good, very good}, and the corresponding quantized data may be {0.01, 0.55, 0.8, 0.91, 1}.

[0077] As an optional embodiment, for the qualitative indicator data, a membership function may be constructed to map each indicator data to a certain membership value on [0, 1] as the quantified indicator data.

[0078] As an optional embodiment, the qualitative indicators may be quantitatively analyzed by constructing a membership function.

[0079] As an optional embodiment, represents the index data after standardization, then the standard quality assessment matrix can be expressed as .

[0080] Step S103: Determine the entropy weight of each quality assessment indicator based on the standard quality assessment matrix and the entropy weight method, and weight the standard quality assessment matrix based on the entropy weight to obtain a weighted standard matrix.

[0081] It should be noted that in information theory, entropy is a measure of uncertainty. The greater the uncertainty, the greater the entropy, and the greater the amount of information contained. The smaller the uncertainty, the smaller the entropy, and the smaller the amount of information contained. According to the characteristics of entropy, the randomness and disorder of an event can be judged by calculating the entropy value, and the discreteness of an indicator can also be judged by the entropy value. The greater the discreteness of the indicator, the greater the influence (weight) of the indicator on the comprehensive evaluation. For example, if the values ​​of a certain sample data under a certain indicator are equal, then the influence (contribution) of the indicator on the overall evaluation is 0, and its weight should also be 0. The entropy weight method is an objective weighting method because it only depends on the discreteness of the data itself. The entropy weight method objectively determines the weight through the distribution of indicator data and information entropy. Unlike the subjective weighting method, the entropy weight method does not rely on human judgment or experience, but determines the weight based on the discreteness and information content of the data itself. The entropy weight method can ensure the objectivity, rationality and dynamic adaptability of weight distribution, highlight important indicators, reduce redundancy, and improve the accuracy and stability of quality assessment. The steps for calculating entropy weight by the entropy weight method are divided into the following four steps: (1) Calculate the probability matrix P .

[0082] (2) Calculate the information entropy of each indicator .

[0083] (3) Calculation of information redundancy .

[0084] (4) Calculating entropy weight .

[0085] As an optional embodiment, the standard quality assessment matrix For example ( ), calculate the The data sample The formula for the weight of the indicator data of a quality assessment indicator is: in, Indicates The data sample The weight of the indicator data of the quality assessment indicator, Indicates The data sample The indicator data of the quality assessment indicators.

[0086] Furthermore, calculate the The entropy value of the quality assessment index , the calculation formula is: in, Represents a data sample, represents the quality assessment index, Indicates The data sample The weight of the indicator data of each quality assessment indicator.

[0087] Further calculate the redundancy of information entropy , . Indicates The entropy value of the quality assessment index Then based on the redundancy Calculating Entropy Weight , the calculation formula is: in, Represents the total number of data samples.

[0088] After obtaining the entropy weight, the entropy weight based on the quality assessment index is multiplied by the standard quality assessment matrix to obtain the weighted standard matrix. The calculation formula is: in, represents the weighted standard matrix, represents the standard quality assessment matrix, Indicates multiplication of corresponding elements of matrices of the same type.

[0089] Step S104: Based on the weighted standard matrix and the ideal solution, the relative proximity of the data samples in the target data set is determined, and the quality assessment result of the target data set is determined based on the relative proximity.

[0090] As an optional embodiment, the ideal solution method is also called TOPSIS method. n evaluation index n coordinate axes, from which we can construct a n dimensional space, then each data sample corresponds to nA coordinate point in a dimensional space. For each evaluation index, select the optimal value (positive ideal solution, corresponding to the optimal coordinate point) and the worst value (negative ideal solution, corresponding to the worst coordinate point) of the index from all data samples, and calculate the Euclidean distance from the coordinate point of each data sample to the optimal coordinate point in turn. The Euclidean distance to the worst coordinate point .

[0091] in, Represents a data sample, represents the quality assessment index, Indicates The negative ideal solution of the quality assessment index, Indicates The positive ideal solution of the quality assessment index is represents the weighted standard matrix The data sample The value of a quality assessment indicator.

[0092] based on and Determining relative proximity .

[0093] As an optional embodiment, the relative proximity is calculated Later, random sampling can be performed to randomly extract the same number of samples from the target data set multiple times, and the relative proximity of the samples of the same data set can be statistically summed and sorted to obtain the comprehensive quality assessment result of the data set. The relative proximity of each data sample can also be summed, and the summed result is used as the comprehensive quality assessment result of the data set. As an optional embodiment, Continue normalization and convert the normalized result As the quality of the sample data of this sampling, the normalization formula is as follows.

[0094] according to It is also possible to judge the pros and cons of each evaluation object. Combined with the steps of determining the indicator weights by the entropy weight method, if the weights of the indicators are the same, the default weights are .in, Represents the total number of data samples.

[0095] As an optional embodiment, for target dataset, for the Sub-sampling, calculate the target data set All The sample data The sum of the sampling data set is obtained. The data quality is expressed as: .

[0096] Sampling multiple times, repeating the above steps, to obtain the data set The comprehensive quality assessment value is recorded as: In the above embodiment, the combination of the two algorithms is achieved based on the comprehensive evaluation of the entropy weight-ideal solution model. The weight of each indicator is calculated by the entropy weight method, and then the comprehensive scoring is performed by the ideal solution analysis method to obtain a quality evaluation result with high credibility.

[0097] As an optional embodiment, the remote sensing data set includes data in the fields of surveying and mapping, meteorology, navigation and time frequency, and a high-quality data set can ensure that the selected data set is more suitable for a specific task. The evaluation method of the present invention not only evaluates the quality of the data samples in the data set, but also can evaluate the overall quality of the data set based on the quality evaluation results of the data samples.

[0098] In the above steps S101-S104, the quality assessment index corresponding to the target data set is determined by the type of the target data set, so that each target data set has its own corresponding quality assessment index, and then the quality assessment matrix is ​​determined based on the index data corresponding to each quality assessment index of the data samples in the target data set, and the quality assessment matrix is ​​preprocessed to obtain a standard quality assessment matrix, eliminating the directional differences and scale differences between the index data. Further, the entropy weight method is used to determine the entropy weight of each quality assessment index in the standard quality assessment matrix to ensure the objectivity of the weight distribution, and the standard quality assessment matrix is ​​weighted based on the entropy weight to obtain a weighted standard matrix. Then, an ideal solution is used for the weighted standard matrix to determine the relative proximity of the data samples in the target data set, and the quality assessment result of the target data set is determined based on the relative proximity, so that the quality assessment result can better reflect the overall quality of the target data set.

[0099] The executor of each step in the above method can be a quality assessment device for massive data sets, which can be implemented through software and / or hardware. The device can be integrated in an electronic device, which can be a terminal device (such as a smart phone, a personal computer, etc.), a server (such as a local server or a cloud server, or a server cluster, etc.), a processor, or a chip, etc.

[0100] Obviously, the above-described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. In order to better understand the above method, the above process is described below in conjunction with the embodiments, but the technical solutions of the embodiments of the present invention are not limited. Specifically: The solution of the present invention is divided into two parts. The first part is to construct a remote sensing big data quality indicator system (corresponding to the quality assessment indicators in the above embodiment). The second part is a quality assessment method for massive data sets based on entropy weight method and ideal solution.

[0101] This application integrates quality features of different dimensions to obtain comprehensive data quality assessment results. Figure 2 This is a quality assessment index system diagram for a massive data set provided by the present invention. Figure 2 As shown in Figure 1, the quality assessment index system includes the overall quality of the data set and the intrinsic quality of the data set. For data sets of the same type, the overall quality index can be used as the quality assessment index of the target data set, the intrinsic quality index can be used as the quality assessment index of the target data set, or the intrinsic quality index and the overall quality index can be used together as the quality assessment index of the target data set. For data sets of different types, since the data content is not comparable, generally only the overall quality index of the data set is selected as the basis for evaluation. After determining the quality assessment index system, the quality of the target data set is evaluated based on the entropy weight method and the ideal solution method. Figure 3 Schematic diagram of the quality assessment method of a massive data set based on the entropy weight method and the ideal solution provided by the present invention. Figure 3 As shown, the method comprises the following steps: (1) Obtain the target data set, determine the type of the target data set, and determine the basis for selecting quality indicators based on whether the target data sets are of the same type.

[0102] (2) Determine the dataset quality indicator system based on the target dataset.

[0103] (3) For each target data set, sample data of the same data is extracted each time.

[0104] (4) Determine the quality assessment matrix.

[0105] (5) Preprocess the indicator data in the quality assessment matrix to obtain the standard quality assessment matrix.

[0106] (6) Use the entropy weight method to weight the standard quality assessment matrix and obtain the weighted standard matrix.

[0107] (7) Determine the optimal distance and the worst distance based on the weighted standard matrix (corresponding to and ).

[0108] (8) Determine the relative proximity through the optimal distance and the worst distance, and obtain the quality assessment result of the data sample sampled this time.

[0109] (9) Calculate the sum of the relative proximity of each sample in each data set in this sampling as the comprehensive quality assessment value of this sampling data set.

[0110] (10) When the required number of sampling times is reached, the sum of the quality assessment values ​​of each data set of K sampling times is calculated and sorted to obtain the comprehensive quality assessment result of the data set.

[0111] The specific calculation process of steps (1)-(10) has been described in detail in the above embodiment and will not be repeated here.

[0112] Aiming at the quality assessment needs of massive data sets, the present invention constructs a quality assessment index system for data sets, comprehensively quantifies indicators of various dimensions, and provides basic data support for the quality assessment of data sets. In addition, a more objective and accurate quality assessment result is obtained by combining the entropy weight method with the ideal solution.

[0113] The quality assessment device for a massive data set provided by the present invention is described below. The quality assessment device for a massive data set described below and the quality assessment method for a massive data set described above can be referred to each other.

[0114] Figure 4 Schematic diagram of the structure of the mass data set quality assessment device provided by the present invention. Figure 4 As shown, the device comprises: The first determination module 410 is used to determine a quality assessment index of a target data set based on the type of the target data set. The type of the target data set includes data sets of the same type and data sets of different types.

[0115] The first processing module 420 is used to determine a quality assessment matrix based on the indicator data corresponding to each quality assessment indicator of the data samples in the target data set, and preprocess the quality assessment matrix to obtain a standard quality assessment matrix.

[0116] The second processing module 430 is used to determine the entropy weight of each quality assessment indicator based on the standard quality assessment matrix and the entropy weight method, and weight the standard quality assessment matrix based on the entropy weight to obtain a weighted standard matrix.

[0117] The second determination module 440 determines the relative proximity of the data samples in the target data set based on the weighted standard matrix and the ideal solution, and determines the quality assessment result of the target data set based on the relative proximity.

[0118] The above-mentioned device determines the quality assessment index corresponding to the target data set by the type of the target data set, so that each target data set has its own corresponding quality assessment index, and then determines the quality assessment matrix based on the index data corresponding to each quality assessment index of the data samples in the target data set, and pre-processes the quality assessment matrix to obtain a standard quality assessment matrix, eliminating the directional differences and scale differences between the index data. Further, the entropy weight method is used to determine the entropy weight of each quality assessment index in the standard quality assessment matrix to ensure the objectivity of the weight distribution, and the standard quality assessment matrix is ​​weighted based on the entropy weight to obtain a weighted standard matrix. Then, an ideal solution is used for the weighted standard matrix to determine the relative proximity of the data samples in the target data set, and the quality assessment result of the target data set is determined based on the relative proximity, so that the quality assessment result can better reflect the overall quality of the target data set.

[0119] In some embodiments, the overall quality indicators include: data completeness, data content diversity, data timeliness, retrieval efficiency, and user satisfaction.

[0120] In some embodiments, the target data set is a data set of remote sensing data, and the remote sensing data includes at least one of digital orthophoto data, digital elevation model data, vector terrain framework data, navigation electronic map data, and real-scene image data.

[0121] In some embodiments, the same type of data set is a data set including any one of digital orthophoto data, digital elevation model data, vector terrain framework data, navigation electronic map data and real-life image data, and the different type of data set is a data set including at least two of digital orthophoto data, digital elevation model data, vector terrain framework data, navigation electronic map data and real-life image data.

[0122] In some embodiments, the method of preprocessing the quality assessment matrix includes at least one of the following: performing uniform processing on indicator data with different indicator directions in the quality assessment matrix. Performing dimensionless processing on indicator data with different dimensions in the quality assessment matrix by standard deviation method, extreme value difference method or efficacy coefficient method. Performing quantization processing on qualitative indicator data in the quality assessment matrix by linear quantization, partial large quantization or membership function.

[0123] Figure 5 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5As shown, the electronic device may include: a processor 510, a communications interface 520, a memory 530 and a communication bus 540, wherein the processor 510, the communications interface 520 and the memory 530 communicate with each other via the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute the quality assessment method of the massive data set, and the method includes: Determining a quality assessment indicator of the target data set based on a type of the target data set, where the type of the target data set includes data sets of the same type and data sets of different types; Determine a quality assessment matrix based on the indicator data corresponding to each of the quality assessment indicators of the data samples in the target data set, and preprocess the quality assessment matrix to obtain a standard quality assessment matrix; Determining the entropy weight of each of the quality assessment indicators based on the standard quality assessment matrix and the entropy weight method, and weighting the standard quality assessment matrix based on the entropy weight to obtain a weighted standard matrix; Based on the weighting standard matrix and the ideal solution, the relative proximity of the data samples in the target data set is determined, and based on the relative proximity, the quality assessment result of the target data set is determined.

[0124] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, 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, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of 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. Various media that can store program codes.

[0125] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the mass data set quality assessment method provided by the above methods, the method includes: Determining a quality assessment indicator of the target data set based on a type of the target data set, where the type of the target data set includes data sets of the same type and data sets of different types; Determine a quality assessment matrix based on the indicator data corresponding to each of the quality assessment indicators of the data samples in the target data set, and preprocess the quality assessment matrix to obtain a standard quality assessment matrix; Determining the entropy weight of each of the quality assessment indicators based on the standard quality assessment matrix and the entropy weight method, and weighting the standard quality assessment matrix based on the entropy weight to obtain a weighted standard matrix; Based on the weighting standard matrix and the ideal solution, the relative proximity of the data samples in the target data set is determined, and based on the relative proximity, the quality assessment result of the target data set is determined.

[0126] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to perform the quality assessment method of the massive data set provided by the above methods, the method comprising: Determining a quality assessment indicator of the target data set based on a type of the target data set, where the type of the target data set includes data sets of the same type and data sets of different types; Determine a quality assessment matrix based on the indicator data corresponding to each of the quality assessment indicators of the data samples in the target data set, and preprocess the quality assessment matrix to obtain a standard quality assessment matrix; Determining the entropy weight of each of the quality assessment indicators based on the standard quality assessment matrix and the entropy weight method, and weighting the standard quality assessment matrix based on the entropy weight to obtain a weighted standard matrix; Based on the weighting standard matrix and the ideal solution, the relative proximity of the data samples in the target data set is determined, and based on the relative proximity, the quality assessment result of the target data set is determined.

[0127] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0128] Through the description of the above implementation modes, those skilled in the art can clearly understand that each implementation mode can be implemented by means of software plus a necessary general hardware platform, or of course by hardware. Based on such an understanding, the above technical solution can essentially or in other words be embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Although the present invention has been described in detail with reference to the above embodiments, a person skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A quality assessment method for a massive data set, characterized in that: include: Determining a quality assessment indicator of the target data set based on a type of the target data set, where the type of the target data set includes data sets of the same type and data sets of different types; Determine a quality assessment matrix based on the indicator data corresponding to each of the quality assessment indicators of the data samples in the target data set, and preprocess the quality assessment matrix to obtain a standard quality assessment matrix; Determining the entropy weight of each of the quality assessment indicators based on the standard quality assessment matrix and the entropy weight method, and weighting the standard quality assessment matrix based on the entropy weight to obtain a weighted standard matrix; Based on the weighting standard matrix and the ideal solution, the relative proximity of the data samples in the target data set is determined, and based on the relative proximity, the quality assessment result of the target data set is determined.

2. The quality assessment method of a massive data set according to claim 1, characterized in that: The determining of the quality assessment index of the target data set based on the type of the target data set includes: In the case where the target data set is a data set of the same type, using the overall quality index and / or the intrinsic quality index of the data set of the same type as the quality assessment index of the target data set; In the case that the target data set is a data set of different types, the overall quality index of the data sets of different types is used as the quality evaluation index of the target data set.

3. The quality assessment method of a massive data set according to claim 2, characterized in that: The overall quality indicators include: data completeness, data content diversity, data timeliness, retrieval efficiency and user satisfaction.

4. The method for quality assessment of a massive data set according to any one of claims 1 to 3, characterized in that: The target data set is a data set of remote sensing data, and the remote sensing data includes at least one of digital orthophoto data, digital elevation model data, vector terrain framework data, navigation electronic map data and real-scene image data.

5. The quality assessment method of a massive data set according to claim 4, characterized in that: The same type of data set is a data set including any one of the digital orthophoto data, the digital elevation model data, the vector terrain framework data, the navigation electronic map data and the real scene image data, and the different type of data set is a data set including at least two of the digital orthophoto data, the digital elevation model data, the vector terrain framework data, the navigation electronic map data and the real scene image data.

6. The quality assessment method of a massive data set according to claim 1, characterized in that: The method of preprocessing the quality assessment matrix includes at least one of the following: Performing consistency processing on indicator data with different indicator directions in the quality assessment matrix; Performing dimensionless processing on the indicator data with different dimensions in the quality assessment matrix by using the standard deviation method, the extreme value difference method or the efficacy coefficient method; The qualitative indicator data in the quality assessment matrix are quantified by linear quantization, partial large-scale quantization or membership function.

7. A quality assessment device for a massive data set, characterized in that: include: A first determination module, configured to determine a quality assessment indicator of a target data set based on a type of the target data set, wherein the type of the target data set includes data sets of the same type and data sets of different types; A first processing module, configured to determine a quality assessment matrix based on the indicator data corresponding to each of the quality assessment indicators of the data samples in the target data set, and preprocess the quality assessment matrix to obtain a standard quality assessment matrix; A second processing module is used to determine the entropy weight of each of the quality assessment indicators based on the standard quality assessment matrix and the entropy weight method, and weight the standard quality assessment matrix based on the entropy weight to obtain a weighted standard matrix; The second determination module is used to determine the relative proximity of the data samples in the target data set based on the weighting standard matrix and the ideal solution, and determine the quality assessment result of the target data set based on the relative proximity.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the quality assessment method for a massive data set as claimed in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the quality assessment method for a massive data set as claimed in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the quality assessment method for a massive data set as claimed in any one of claims 1 to 6 is implemented.

Images