Part quality grade classification method and device based on uncertain data pattern classification, equipment, medium and product

Abstract

The invention discloses a part quality grade classification method and device based on uncertain data pattern classification, equipment, a medium and a product, and relates to the field of artificial intelligence. The method comprises the following steps: firstly, acquiring part size data including multiple dimension characteristics of different parts and different processing stages of a part monitored by a sensor; performing data preprocessing on the part size data to obtain preprocessed part size data; a fitness function of a genetic algorithm is improved based on granularity and an elastic network, and a feature selection genetic algorithm is formed; screening a plurality of dimension features of the preprocessed part size data by using a feature selection genetic algorithm to obtain an optimal feature subset; extracting feature data corresponding to the optimal feature subset from the preprocessed part size data, and constructing a feature data set; based on the feature data set, part quality grade classification is carried out by using the evidence K neighbor model, so that the efficiency and accuracy of part quality grade classification in an uncertain data environment can be remarkably improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, device, equipment, medium and product for classifying part quality grades based on uncertain data pattern classification. Background Art

[0002] In the current context of artificial intelligence and data science, a major challenge is how to effectively process and analyze uncertain data. Uncertain data may come from various scenarios, such as sensor data collection, social media analysis, medical diagnosis, etc., and they often contain noise, missing values, or fuzzy information. Existing uncertain data processing and pattern recognition methods may be inefficient or inaccurate when processing such uncertain data. Therefore, developing new algorithms and models to improve the ability to process uncertain data while maintaining efficiency and accuracy has become an important research direction.

[0003] In the field of industrial manufacturing, part size data is affected by factors such as precision and wear, and there is uncertainty, which is a typical uncertain data. If such uncertain data is directly used to classify the quality grade of parts, on the one hand, the processing efficiency will be low due to too much invalid or redundant data, and on the other hand, the uncertainty of part size data will greatly affect the accuracy of part quality grade classification. Summary of the invention

[0004] The purpose of this application is to provide a method, device, equipment, medium and product for part quality grade classification based on uncertain data pattern classification, so as to improve the efficiency and accuracy of part quality grade classification in an uncertain data environment.

[0005] To achieve the above objectives, this application provides the following solutions.

[0006] In a first aspect, the present application provides a method for classifying part quality grades based on uncertain data pattern classification, comprising: Obtain part dimensional data containing multiple dimensional features at different parts and different processing stages of the parts monitored by sensors; Performing data preprocessing on the part size data to obtain preprocessed part size data; Improve the fitness function of genetic algorithm based on granularity and elastic network to form feature selection genetic algorithm; Use feature selection genetic algorithm to screen multiple dimensional features of preprocessed part size data to obtain the optimal feature subset; Extract feature data corresponding to the optimal feature subset from the preprocessed part size data to construct a feature data set; Based on the feature data set, the evidential K nearest neighbor model is used to classify the quality grades of parts.

[0007] Optionally, the preprocessing of the part size data to obtain the preprocessed part size data specifically includes: De-noising, missing value filling and normalization techniques are used to preprocess the part size data to obtain the preprocessed part size data.

[0008] Optionally, the fitness function of the genetic algorithm is improved based on granularity and elastic network to form a feature selection genetic algorithm, specifically including: In the genetic algorithm, each individual chromosome represents a set of feature selection schemes, and each gene on the chromosome represents whether the corresponding feature is selected. individuals, and the length of the gene code of each individual is ,use To indicate the The chromosomes of an individual, express No. Gene loci; ; The entire population is represented by ; Assumptions is the population of the current generation, Contains all genes of the current generation, and Represents all the selected genes in the current generation population, according to the formula Calculate the granularity of the current generation of individuals ;in The current chromosome The selected feature subset in , is the number of chromosome entries selected; is a feature subset The number of features selected in ; is the total number of genes in the current generation; Elastic Net calculates prediction error on fold data via cross validation ;in It is a test sample The true label of It is a test sample The predicted label of is the number of test samples; Based on the granularity of the current generation and prediction error Constructing an improved fitness function based on granularity and elastic network ;in is the weight factor; is the calculated fitness value; The improved fitness function is used as the fitness function of the genetic algorithm to form a feature selection genetic algorithm.

[0009] Optionally, the use of a feature selection genetic algorithm to screen multiple dimensional features of the preprocessed part size data to obtain an optimal feature subset specifically includes: The initial population of the feature selection genetic algorithm is randomly generated. The chromosome of each individual in the initial population represents a set of feature selection schemes. Each gene on the chromosome represents whether the corresponding feature is selected. The gene encoding length of each individual is equal to the dimension of the features contained in the preprocessed part size data. The iterative calculation of the genetic algorithm is performed from the initial population. For each individual in each generation of the population, the fitness value is calculated using the improved fitness function until the predetermined number of iterations is reached; After reaching the predetermined number of iterations, the feature subset represented by the individual chromosome with the highest fitness value is output as the optimal feature subset.

[0010] Optionally, extracting feature data corresponding to the optimal feature subset from the preprocessed part size data to construct a feature data set specifically includes: Each feature has a corresponding index position in the optimal feature subset. In the preprocessed part size data, the data of the corresponding feature column is located and extracted as the corresponding feature data according to the index position corresponding to the feature; all feature data extracted from all preprocessed part size data together constitute a feature data set.

[0011] Optionally, the method of classifying the quality grade of parts based on the feature data set using the evidence K nearest neighbor model specifically includes: The feature data set is divided into a training set and a test set. The training set is used to train the evidence K nearest neighbor model, and the test set is used to test the prediction error of the evidence K nearest neighbor model based on the optimal feature subset. ; After the training is completed, the trained evidence K nearest neighbor model is used as the part quality grade classification model; The feature vector of the part size data to be classified is input into the part quality grade classification model, and the corresponding part quality grade is output; the part quality grade includes qualified, slightly unqualified and seriously unqualified.

[0012] In a second aspect, the present application provides a device for classifying part quality grades based on uncertain data pattern classification, comprising: The uncertain data acquisition module is used to obtain the part size data containing multiple dimensional features at different parts and different processing stages of the parts monitored by sensors; A data preprocessing module is used to preprocess the part size data to obtain the preprocessed part size data; The fitness function improvement module is used to improve the fitness function of the genetic algorithm based on granularity and elastic network to form a feature selection genetic algorithm; A feature screening module is used to screen multiple dimensional features of preprocessed part size data using a feature selection genetic algorithm to obtain an optimal feature subset; A feature data extraction module is used to extract feature data corresponding to the optimal feature subset from the preprocessed part size data and construct a feature data set; The part quality grade classification module is used to classify the part quality grade based on the feature data set using the evidence K nearest neighbor model.

[0013] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for classifying part quality grades based on uncertain data pattern classification.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for classifying part quality grades based on uncertain data pattern classification.

[0015] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the method for classifying part quality grades based on uncertain data pattern classification.

[0016] According to the specific embodiments provided in this application, this application discloses the following technical effects.

[0017] The present application provides a method, device, equipment, medium and product for part quality grade classification based on uncertain data pattern classification. The fitness function of the genetic algorithm is improved based on granularity and elastic network, and the data features that affect the classification results are optimized and selected through the genetic algorithm to extract the feature data corresponding to the optimal feature subset; the selected feature data is input into the evidence K nearest neighbor model, and the uncertainty is quantified using the trust function, so that the uncertain data is more effectively classified, which can significantly improve the efficiency and accuracy of part quality grade classification in an uncertain data environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 This is a flow chart of a method for classifying part quality grades based on uncertain data pattern classification in the present application. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0021] In today's data-intensive application scenarios, the uncertainty and incompleteness of data have become a problem that cannot be ignored. The existence of these problems significantly affects the accuracy and reliability of data analysis. In order to overcome this problem, this application proposes a method, device, equipment, medium and product for part quality grade classification based on uncertain data pattern classification, aiming to optimize feature selection and process uncertainty information by combining an improved feature selection genetic algorithm and an evidence K nearest neighbor model, so as to achieve more accurate data classification and improve the efficiency and accuracy of part quality grade classification in an uncertain data environment.

[0022] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0023] In an exemplary embodiment, Figure 1 As shown, a method for classifying part quality grades based on uncertain data pattern classification is provided, including the following steps 1 to 6.

[0024] Step 1: Obtain part dimensional data containing multiple dimensional features at different locations and processing stages of the part monitored by sensors.

[0025] In the field of industrial manufacturing, part dimensional data is subject to uncertainty due to factors such as precision and wear. By using sensors or other data acquisition methods to monitor the state of parts, we can obtain part dimensional data containing multiple dimensional features at different parts and processing stages of the parts. The multiple dimensional features of part dimensional data can be, for example, the numerical range features of part dimensional data, the change features of different parts, the change features of different processing stages and at different times, and so on.

[0026] Step 2: Preprocess the part size data to obtain preprocessed part size data.

[0027] The original part size data obtained by sensors or other data collection methods may contain noise, incompleteness or ambiguity, resulting in uncertainty. In the data preprocessing stage, denoising, missing value filling, normalization and other technologies are applied to reduce the uncertainty in the data and obtain the preprocessed part size data. The preprocessed part size data is used in the subsequent feature screening and model training and testing processes, and will no longer be distinguished from the original part size data.

[0028] Step 3: Improve the fitness function of the genetic algorithm based on granularity and elastic network to form a feature selection genetic algorithm.

[0029] Genetic Algorithm (GA) is a search algorithm that simulates natural selection and genetics mechanisms to solve optimization problems by simulating the "survival of the fittest" principle. The main steps of the genetic algorithm are as follows S1 to S6.

[0030] S1 Population initialization: Randomly generate the initial population. The chromosome of each individual in the population represents a set of feature selection schemes, and each gene on each chromosome represents whether the corresponding feature is selected. In the genetic algorithm, each individual is represented by a string of binary codes, which is called a chromosome. Assume that there are individuals, and the length of the gene code of each individual is .use To indicate the The chromosomes of an individual can also be called individuals .in Represents an individual No. Gene loci, is a binary value, which can be either 0 or 1, for example 0 means the corresponding feature is not selected. 1 indicates that the corresponding feature is selected. ; . This encoding process can be expressed as: (1); The entire population can be expressed as: (2).

[0031] S2 Fitness calculation: For each individual in the population, its performance is evaluated by the fitness function. The fitness function is used to evaluate the performance of each individual in the current generation. Individuals with high fitness values ​​are more likely to be selected as "excellent parents" and retained for mating to generate the next generation. The design of the fitness function should be closely related to the goal of feature selection. This application improves the fitness function of the genetic algorithm based on granularity and elastic network. The improved fitness function comprehensively considers the classification accuracy and the granularity information of feature selection.

[0032] S3 selection operation: Based on the calculated fitness value, roulette or other selection mechanisms are used to select individuals with high fitness to be inherited to the next generation, ensuring the inheritance of excellent genes in the population.

[0033] S4 Crossover and mutation. Crossover: Randomly pair individuals according to the crossover probability. Perform crossover operations to produce new offspring. Mutation: Genes of individuals are modified according to the probability of mutation. Perform random mutations to increase the diversity of the population.

[0034] S5 Formation of a new generation population: After selection, crossover and mutation operations, a new generation population is formed for the next round of iteration.

[0035] S6 algorithm iteration: Repeat S2 to S5 until the predetermined number of iterations is reached At this time, the algorithm is considered to have converged to the optimal or approximately optimal solution, and the individual with the highest fitness value is output, which is called the optimal individual.

[0036] Ultimately, the genetic algorithm outputs the feature subset represented by the individual (chromosome) with the highest fitness value as the optimal feature subset for subsequent data analysis or pattern recognition tasks.

[0037] The step 3 specifically includes the following steps 3.1 to 3.5.

[0038] Step 3.1: Use the chromosome of each individual in the genetic algorithm to represent a set of feature selection schemes, and each gene on the chromosome represents whether the corresponding feature is selected; assuming that there are individuals, and the length of the gene code of each individual is ,use To indicate the The chromosomes of an individual, express No. Gene loci; ; The entire population is represented by .

[0039] Step 3.2: Assumptions is the population of the current generation, Contains all genes of the current generation, and Represents all the selected genes in the current generation population, according to the formula Calculate the granularity of the current generation of individuals ;in The current chromosome The selected feature subset in , is the number of chromosome entries selected; is a feature subset The number of features selected in ; is the total number of genes in the current generation.

[0040] Granularity is used to measure the coverage of feature space by feature subsets and control the number and diversity of selected features. Its main functions are as follows: ① Control the size of feature subsets: Granularity encourages the selection of relatively small feature subsets, which can avoid selecting too many features, thereby reducing the risk of feature redundancy and overfitting. Through the introduction of granularity, genetic algorithms can tend to select more compact and efficient feature subsets. ② Promote effective feature selection: Granularity guides genetic algorithms to select subsets that can effectively cover the feature space to the greatest extent in each generation by measuring the "sparseness" or "density" of feature subsets. Therefore, granularity can help genetic algorithms select features that can fully capture data changes, thereby improving the performance of the model. ③ Improve convergence efficiency: In genetic algorithms, granularity helps accelerate the convergence process and reduce the selection of invalid or redundant features. By gradually optimizing the granularity, genetic algorithms can effectively find the optimal feature subset and accelerate convergence to the optimal solution.

[0041] Specifically, assuming is the population of the current generation, Contains all genes of the current generation, and Represents all the selected genes in the current generation population, where The current chromosome The selected feature subset in , is the number of selected chromosome entries. Then the granularity of the current generation of individuals Defined as: (3); in, is a feature subset The number of features selected in ; is the total number of genes in the current generation. If , then it means Chromosomes The fewest features are selected; if , then it means Chromosomes All features are selected.

[0042] Step 3.3: Elastic Net calculates the prediction error of the fold data through cross validation ;in It is a test sample The true label of It is a test sample The predicted label of is the number of test samples.

[0043] Elastic Net (Elastic Net) regularization combines L1 (Lasso) and L2 (Ridge) regularization and is widely used in regression problems, especially feature selection in high-dimensional data. Its main functions are as follows: 1) Improving the effectiveness of feature selection: Elastic Net uses L1 regularization to make the coefficients of some features zero, thereby achieving feature selection. L2 regularization helps control the size of the regression coefficient to prevent overfitting. The combination of the two enables Elastic Net to find a balance between feature selection and regularization, removing unimportant features and retaining features that are useful to the model. 2) Adapting to high-dimensional data: In high-dimensional data, the number of features may be much larger than the number of samples. Elastic Net can effectively perform feature selection and avoid the problem of overfitting of the model due to high data dimensions. By improving the fitness function through the elastic net, the genetic algorithm can screen out the most valuable features in the high-dimensional feature space and improve the generalization ability of the model.

[0044] The elastic network is defined as follows. Assume that there is a linear regression polynomial model ,in is the response variable (target), is the design matrix containing all feature variables, is the coefficient vector to be estimated, is the error term. Assume that common predictor variables, denoted by Based on linear regression, the response variable The estimate is expressed as The coefficient vector It is calculated by minimizing the sum of squared errors. The objective function of the elastic net can be written as: (4); in, (5); (6); In the formula, Express request Minimum value of the intrinsic function. The first item It is used to minimize the sum of squared errors of the fitted data. represents the L1 norm, which promotes the sparsity of coefficients and encourages some coefficients to become zero, thereby achieving feature selection. Represents the L2 norm, which controls the sum of squares of coefficients to prevent overfitting.

[0045] Elastic Net calculates prediction error on fold data via cross validation , evaluate the performance of the algorithm.

[0046] (7); in It is a test sample The true label of It is a test sample The predicted label of is the number of test samples.

[0047] Step 3.4: Granularity based on the current generation of individuals and prediction error Constructing an improved fitness function based on granularity and elastic network ;in is the weight factor; Calculated fitness value.

[0048] The improved fitness function based on granularity and elastic network is set as: (8); in, The model prediction error value obtained for the feature subset corresponding to the genetic code of individuals in the current generation; Indicates the granularity of the current generation of individuals. Since the granularity is expected to become larger and larger and will limit the impact on the fitness function, a weight factor is added in front ,generally = 0.15. This value is not arbitrary, but rather it was determined through multiple iterations and experimental validation to ensure that the granularity contributes appropriately to the overall fitness assessment, balancing and the impact of feature subset size on the performance of genetic algorithms.

[0049] The feature selection of the genetic algorithm uses the improved fitness function (8) to identify the best individual in each iteration. The elastic network removes redundant features and retains the most valuable features. Features that frequently appear in the best individuals are selected as the final features and constitute the best feature subset. This process aims to reduce And increase the granularity of iterations.

[0050] Step 3.5: Use the improved fitness function as the fitness function of the genetic algorithm to form a feature selection genetic algorithm.

[0051] Given the improved fitness function (8), the optimization problem of the genetic algorithm is to find the optimal chromosome, that is, the optimal individual , that is, the combination of the optimal features, is called the optimal feature subset. satisfy: (9); in Function representation makes the objective function The value of the variable when it reaches its minimum value.

[0052] Step 4: Use feature selection genetic algorithm to screen multiple dimensional features of preprocessed part size data to obtain the optimal feature subset.

[0053] For the part size data obtained from the industrial manufacturing field, the population initialization step in the genetic algorithm directly relies on these uncertain data. After being collected and preprocessed by sensors, the part size data becomes the basic material of the genetic algorithm. The feature dimension corresponding to each part size data is like the gene position on a chromosome. For example, different measurement dimensions of part size data are encoded into binary values, which are combined to form individuals representing feature selection schemes, and many such individuals constitute the initial population. This means that uncertain data provides the original feature resources for the genetic algorithm, giving it an operational object, which is the foundation for the subsequent algorithm process to start.

[0054] First, the candidate feature subsets are generated by initializing the population, and then the improved fitness function is used to evaluate the contribution of each feature subset to the classification task. The improved fitness function not only considers the classification performance of the feature subset, but also introduces the root mean square error (RMSE) to measure the error of the model, and combines it with the elastic network for analysis. The elastic network can effectively reduce the weight of useless features by combining L1 and L2 regularization on the features, thereby reducing the error and improving the generalization ability of the model. On this basis, through genetic operations such as crossover and mutation, the feature subsets are iteratively optimized, and finally the most representative optimal feature subset is selected. The granularity benchmark mechanism quantifies the influence of each feature through the mathematical formula of the combined granularity, providing a quantitative evaluation standard to guide the selection process in the genetic algorithm.

[0055] The step 4 specifically includes: Step 4.1: Randomly generate the initial population of the feature selection genetic algorithm. The chromosome of each individual in the initial population represents a set of feature selection schemes. Each gene on the chromosome represents whether the corresponding feature is selected. The gene encoding length of each individual is equal to the dimension of the features contained in the preprocessed part size data. Step 4.2: Starting from the initial population, perform iterative calculation of the genetic algorithm. For each individual in each generation of the population, use the improved fitness function (8) to calculate the fitness value until the predetermined number of iterations is reached; Step 4.3: After reaching the predetermined number of iterations, the feature subset represented by the individual chromosome with the highest fitness value is output as the optimal feature subset.

[0056] This application adopts an improved binary genetic algorithm and feature granulation strategy to measure candidate features through a granularity benchmark mechanism; the improved fitness function combines classification accuracy and feature granularity, greatly optimizing the feature selection process.

[0057] Step 5: Extract feature data corresponding to the optimal feature subset from the preprocessed part size data and construct a feature data set.

[0058] After the optimal feature subset is selected using the feature selection genetic algorithm, each feature has a corresponding index position in the optimal feature subset. In the preprocessed part size data, according to the index position corresponding to the feature, the data of the corresponding feature column can be accurately located and extracted as the corresponding feature data. All feature data extracted from all preprocessed part size data together constitute the feature data set.

[0059] For example, all preprocessed part size data constitute the original data set. If the optimal feature subset contains 5 features, corresponding to the 3rd, 7th, 11th, 15th, and 19th columns of features in the original data set, then these 5 columns of data are extracted from each row of data in the original data set to form a new data set containing only the feature data corresponding to the optimal feature subset, which is called a feature data set. This method is simple and direct, and can quickly complete data extraction according to established rules. It is especially suitable for structured data and can ensure the accuracy and efficiency of data extraction.

[0060] Subsequently, the feature data in the feature data set are divided into a training set and a test set according to a certain ratio, and the feature subset represented by the individual is applied to the evidence K nearest neighbor model. After training with the training set, the classification accuracy of the model based on the feature subset is tested based on the test set. This accuracy is a key component of the fitness function, which measures whether the feature subset can effectively deal with the classification problem caused by data uncertainty. On the other hand, the granularity information extracted from the uncertain data is integrated into the fitness function. The granularity benchmark mechanism is based on data characteristics, such as the fluctuation granularity of part size data at different processing stages. The granularity of the selected features in the chromosome is quantified by formula (3), and then the influence of each feature is quantified, providing an additional evaluation dimension for the fitness function, allowing it to more accurately screen out feature subsets that can effectively deal with uncertainty.

[0061] Step 6: Based on the feature data set, use the evidence K nearest neighbor model to classify the part quality grade.

[0062] In this application, the K-nearest neighbor model based on evidence theory is used to deal with uncertainty, referred to as the Evidential K-Nearest Neighbors (EKNN) model. The EKNN model is used to realize the use of uncertain information of data and pattern recognition and classification. If the category set of uncertain data Is The uncertain data is composed of mutually exclusive elements, which is called the pattern recognition framework of uncertain data, also known as the identification framework. .

[0063] Evidence theory provides the Dempster combination rule, which can perform orthogonal combination of multiple information and realize the fusion of multiple classification evidence. , Etc. usually represent different focal elements, that is, subsets in the problem domain under study. Taking the application of the EKNN model in classification tasks as an example, suppose that the quality level of parts in the industrial manufacturing field is to be classified. The full set here is the set of all possible quality levels of parts, such as {qualified, slightly unqualified, seriously unqualified}. can represent the subset "qualified", It can represent the subset of "minor failures", and so on.

[0064] and They are defined in the same identification framework There are two basic probability assignment functions on , and It represents the probability of different focal elements being generated under the probability assignment function. The calculation formula of Dempster combination rule is as follows: (10); in The evidence and The new classification evidence obtained by integration; Represents a new focus element. It means that the probability of the empty set is assigned a value of 0, which can be expressed as: (11); is the conflict coefficient used to measure the degree of conflict between evidences. The larger the value, the greater the conflict between the evidence. =1, the combination rule cannot be used. Very large, indicating a high level of conflict. The choice of value will affect the accuracy of classification. The optimal value can be determined by cross-validation and other methods. value.

[0065] In this application, the trust membership of EKNN (referred to as trust) is mapped using Euclidean distance. In the distance-based trust allocation algorithm, the relationship between each scattered data point and The distance between the classification center points is obtained Now assume a basic credit distribution function , assuming that any two-dimensional scattered point is , calculate the scattered points to The distance between cluster centers , normalize all distance values. Will Mapping probability values, each with express, . Two conditions need to be met: , and the sum of all distribution reliability results is 1; second, the distribution function should be a monotonically decreasing function, that is, as the distance from the scattered points to the classification center point increases, the reliability distribution decreases.

[0066] The step 6 specifically includes the following steps 6.1 and 6.2.

[0067] Step 6.1: Divide the feature data set into a training set and a test set, use the training set to train the evidence K nearest neighbor model, and use the test set to test the prediction error of the evidence K nearest neighbor model based on the optimal feature subset. ; After the training is completed, the trained evidence K nearest neighbor model is used as the part quality grade classification model.

[0068] Assume a group The training set consists of feature data .in is a class label The characteristic vector of is the dimensional space of feature vectors; In this application, the obtained feature data is organized to conform to the input format of the EKNN model, which includes extracting and organizing each part size data sample according to the feature dimension determined by the optimal feature subset to obtain feature data, and then composing it into a feature vector. For example, if the optimal feature subset contains three features, then each feature data is a feature vector composed of the values ​​of these three features. The feature data is represented in the form of a concatenated vector, so it can also be called a feature vector. Assume is the feature vector to be classified, which comes from the test set and is called the test vector here. yes exist In The set of nearest neighbors. Every neighbor of , whose label exist In the A clear set of evidence for computing The category to which it belongs.

[0069] This evidence set can be calculated using the following distance mapping function: (12); (13); in is a constant such that and is a positive parameter, usually related to the class Related. Usually fixed And for all , . Indicates that in the known The sample is assumed to belong to the category The probability assignment of Indicates that in the known The probability assignment of all possible categories to which the sample belongs under the premise.

[0070] if and Very "close", so believe and belong to the same category; otherwise, There is a lot of uncertainty about the category. Class, using Dempster's combination rule, all Evidence items Can be combined as: (14); Represents the fusion result of all evidences. Symbol Indicates finding the intersection of the evidence information involved.

[0071] Finally, the category decision is made through the fused evidence, and the category with the highest probability is the The category to which the characteristic data of the sample belongs, that is, qualified, slightly unqualified or seriously unqualified.

[0072] Step 6.2: Input the feature vector of the part size data to be classified into the part quality grade classification model, and output the corresponding part quality grade; the part quality grade includes qualified, slightly unqualified and seriously unqualified.

[0073] This application uses a genetic algorithm combined with granularity and elastic network regularization to perform feature selection. In multiple iterations, the genetic algorithm selects the optimal feature subset based on an adaptive function based on granularity and elastic network improvement. These optimal feature subsets can retain the key information of the data to the greatest extent and remove redundant features, thereby optimizing subsequent classification tasks. Feature data corresponding to the optimal feature subset is extracted from the part size data to be classified, and input into the trained part quality grade classification model in the form of a feature vector for pattern recognition and classification, and the corresponding part quality grade can be output.

[0074] This application measures candidate features through a granularity benchmark mechanism, utilizes an improved fitness function that balances classification accuracy and granularity, uses a genetic algorithm to optimize the selection of features of input data, and identifies the data features that have the greatest impact on pattern classification by simulating the process of natural selection; the screened feature data is input into the EKNN model, and the conflicts between evidences are quantified using the DS evidence theory; the innovative conflict quantification allocation method of this application can effectively process and utilize uncertainty information, has a wide range of applicability, can effectively process uncertain data from various sources, and improve the accuracy and efficiency of pattern recognition.

[0075] Therefore, the method of the present application is not only applicable to the analysis and processing of part size data, but can also be extended to process uncertain data in other fields. For example, the temperature and humidity data of the production environment are affected by many factors in the workshop and are uncertain, and are collected by temperature and humidity sensors; in the field of transportation, vehicle speed data is affected by road conditions, vehicle performance and external interference, and there is uncertainty, and vehicle-mounted sensors are used to obtain uncertain data. The features corresponding to each physical quantity in the uncertain data (for example, different dimensional features corresponding to data such as temperature and wind speed in the meteorological field, such as the numerical range features of temperature, the characteristics of wind speed changes in different time periods, etc.) will become the basic elements for constructing a candidate feature set. These original uncertain data features are combined and arranged to form a large number of candidate feature sets in the initialization population. In other words, the various feature variables contained in the uncertain data become specific objects that can be selected and combined when the population is initialized, thereby generating many different feature subset combinations as the starting point for the subsequent iterative optimization of the feature selection genetic algorithm.

[0076] Taking the field of industrial manufacturing as an example, for part size data, a set of data containing multiple dimensional feature values ​​obtained by measuring different parts of the parts and different processing stages each time is an original data sample; in terms of production environment temperature and humidity data, the combination of temperature and humidity values ​​collected in a specific time period and a specific workshop location is also an original data sample. Similarly, in the field of transportation, the speed value of a vehicle at a certain moment and its associated road conditions, vehicle status and other feature values ​​together constitute an original data sample. Applying the method of this application to the field of industrial manufacturing can classify the quality level of parts into qualified, slightly unqualified, seriously unqualified, etc. When applied to the classification of production environment status, it can be judged whether the environment is suitable for production, temperature and humidity abnormalities need to be adjusted, etc. When applied to the classification of vehicle safety status, it can be evaluated whether the vehicle is currently safe or whether there are safety hazards of varying degrees (mild, moderate, severe), etc.

[0077] In an exemplary embodiment, the present application also provides a device for classifying part quality grades based on uncertain data pattern classification, comprising: The uncertain data acquisition module is used to obtain the part size data containing multiple dimensional features at different parts and different processing stages of the parts monitored by sensors; A data preprocessing module is used to preprocess the part size data to obtain the preprocessed part size data; The fitness function improvement module is used to improve the fitness function of the genetic algorithm based on granularity and elastic network to form a feature selection genetic algorithm; A feature screening module is used to screen multiple dimensional features of preprocessed part size data using a feature selection genetic algorithm to obtain an optimal feature subset; A feature data extraction module is used to extract feature data corresponding to the optimal feature subset from the preprocessed part size data and construct a feature data set; The part quality grade classification module is used to classify the part quality grade based on the feature data set using the evidence K nearest neighbor model.

[0078] This application uses a genetic algorithm to optimize the features of the input data, and identifies the data features that have the greatest impact on pattern classification by simulating the process of natural selection. The candidate features are measured by a granularity benchmark mechanism, and a mathematical formula for granularity is defined to quantify the granularity of the selected features in the chromosome, and to optimize the selection of data features that affect the classification results. The accuracy and efficiency of pattern recognition are improved by balancing the improved fitness function of classification accuracy and granularity. Then, the screened feature data is input into an EKNN model, and the uncertainty is quantified and processed using a trust function to achieve more accurate pattern recognition and classification. The EKNN model uses DS evidence theory to deal with uncertainty and ambiguity, and optimizes the application of trust functions in the classification process by quantifying and resolving conflicts between evidence. This method allows the model to make decisions and classifications more effectively when faced with incomplete or inconsistent information. Through an innovative conflict quantification allocation method, the EKNN model can effectively process and utilize uncertainty information without sacrificing classification accuracy, further improving the performance and reliability of the uncertain data pattern classification method based on trust function theory. This innovation not only enhances the robustness of the model, but also provides a new tool for processing pattern recognition in complex data environments.

[0079] In an exemplary embodiment, the present application also provides a computer device, which may be a server or a terminal. The computer device includes a processor, a memory, an input / output interface, and a communication interface. The processor, the memory, and the input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the part quality grade classification method based on uncertain data pattern classification is implemented.

[0080] In an exemplary embodiment, the present application also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for classifying part quality grades based on uncertain data pattern classification is implemented.

[0081] In an exemplary embodiment, the present application also provides a computer program product, including a computer program, which implements the part quality grade classification method based on uncertain data pattern classification when executed by a processor.

[0082] It can be understood by a person skilled in the art that all or part of the processes in the above-mentioned embodiment method can be completed by hardware related to computer program instructions, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the process of the embodiment of the above-mentioned method. Among them, any reference to the memory or other medium in each embodiment provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0083] It should be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0084] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0085] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for classifying parts quality grades based on uncertain data pattern classification, characterized in that: include: Obtain part dimensional data containing multiple dimensional features at different parts and different processing stages of the parts monitored by sensors; Performing data preprocessing on the part size data to obtain preprocessed part size data; Improve the fitness function of genetic algorithm based on granularity and elastic network to form feature selection genetic algorithm; Use feature selection genetic algorithm to screen multiple dimensional features of preprocessed part size data to obtain the optimal feature subset; Extract feature data corresponding to the optimal feature subset from the preprocessed part size data to construct a feature data set; Based on the feature data set, the evidential K nearest neighbor model is used to classify the quality grades of parts.

2. The method for classifying parts quality grades based on uncertain data pattern classification according to claim 1 is characterized in that: The preprocessing of the part size data to obtain the preprocessed part size data specifically includes: De-noising, missing value filling and normalization techniques are used to preprocess the part size data to obtain the preprocessed part size data.

3. The method for classifying parts quality grades based on uncertain data pattern classification according to claim 1, characterized in that: The fitness function of the genetic algorithm is improved based on granularity and elastic network to form a feature selection genetic algorithm, which specifically includes: In the genetic algorithm, each individual chromosome represents a set of feature selection schemes, and each gene on the chromosome represents whether the corresponding feature is selected. individuals, and the length of the gene code of each individual is ,use To express the The chromosomes of an individual, express No. Gene loci; ; The entire population is represented by ; Assumptions is the population of the current generation, Contains all genes of the current generation, and Represents all the selected genes in the current generation population, according to the formula Calculate the granularity of the current generation of individuals ;in The current chromosome The selected feature subset in , is the number of chromosome entries selected; is a feature subset The number of features selected in ; is the total number of genes in the current generation; Elastic Net calculates prediction error on fold data via cross validation ;in It is a test sample The true label of It is a test sample The predicted label of is the number of test samples; Based on the granularity of the current generation and prediction error Constructing an improved fitness function based on granularity and elastic network ;in is the weight factor; The fitness value calculated for the current generation; The improved fitness function is used as the fitness function of the genetic algorithm to form a feature selection genetic algorithm.

4. The method for classifying parts quality grades based on uncertain data pattern classification according to claim 3 is characterized in that: The method uses a feature selection genetic algorithm to screen multiple dimensional features of the preprocessed part size data to obtain an optimal feature subset, specifically including: The initial population of the feature selection genetic algorithm is randomly generated. The chromosome of each individual in the initial population represents a set of feature selection schemes. Each gene on the chromosome represents whether the corresponding feature is selected. The gene encoding length of each individual is equal to the dimension of the features contained in the preprocessed part size data. The iterative calculation of the genetic algorithm is performed from the initial population. For each individual in each generation of the population, the fitness value is calculated using the improved fitness function until the predetermined number of iterations is reached; After reaching the predetermined number of iterations, the feature subset represented by the individual chromosome with the highest fitness value is output as the optimal feature subset.

5. The method for classifying parts quality grades based on uncertain data pattern classification according to claim 4 is characterized in that: The extracting feature data corresponding to the optimal feature subset from the preprocessed part size data and constructing a feature data set specifically includes: Each feature has a corresponding index position in the optimal feature subset. In the preprocessed part size data, the data of the corresponding feature column is located and extracted as the corresponding feature data according to the index position corresponding to the feature; all feature data extracted from all preprocessed part size data together constitute a feature data set.

6. The method for classifying parts quality grades based on uncertain data pattern classification according to claim 5 is characterized in that: Based on the feature data set, the evidence K nearest neighbor model is used to classify the quality level of the parts, specifically including: The feature data set is divided into a training set and a test set. The training set is used to train the evidence K nearest neighbor model, and the test set is used to test the prediction error of the evidence K nearest neighbor model based on the optimal feature subset. ; After the training is completed, the trained evidence K nearest neighbor model is used as the part quality grade classification model; The feature vector of the part size data to be classified is input into the part quality grade classification model, and the corresponding part quality grade is output; the part quality grade includes qualified, slightly unqualified and seriously unqualified.

7. A device for classifying parts quality grades based on uncertain data pattern classification, characterized in that: include: The uncertain data acquisition module is used to obtain the part size data containing multiple dimensional features at different parts and different processing stages of the parts monitored by sensors; A data preprocessing module is used to preprocess the part size data to obtain the preprocessed part size data; The fitness function improvement module is used to improve the fitness function of the genetic algorithm based on granularity and elastic network to form a feature selection genetic algorithm; A feature screening module is used to screen multiple dimensional features of preprocessed part size data using a feature selection genetic algorithm to obtain an optimal feature subset; A feature data extraction module is used to extract feature data corresponding to the optimal feature subset from the preprocessed part size data and construct a feature data set; The part quality grade classification module is used to classify the part quality grade based on the feature data set using the evidence K nearest neighbor model.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for classifying part quality grades based on uncertain data pattern classification according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for classifying part quality grades based on uncertain data pattern classification described 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 method for classifying part quality grades based on uncertain data pattern classification described in any one of claims 1 to 6 is implemented.

Images