Multi-view dictionary learning-based classification method for mixed sampling industrial big data

Abstract

The invention discloses a multi-view dictionary learning-based classification method for mixed sampling industrial big data. According to the method, the classification idea of the multi-view data isingeniously utilized, so that the connection is carried out according to the common characteristics of the mixed sampling industrial data and the multi-view data; Meanwhile, the classification schemeadapting to the mixed sampling data is designed by considering the characteristic that the number of samples of the mixed sampling data is inconsistent. At a training stage, a dictionary of each typeof training samples of each sampling frequency data is learned through a dictionary learning method. At a classification test stage, firstly, a trained dictionary related to the corresponding samplingfrequency data is utilized to encode a test sample, then the minimum reconstruction error between the sample and which type of cluster is judged by utilizing the sub-dictionary and the coding vectorof the test sample, so that the sample is indicated to belong to the type of cluster. Compared with the prior art, and according to the method, the original data is utilized to the maximum extent, thedistribution of the original data is ensured, and the classification result precision is improved.

Description

technical field [0001] The invention belongs to the technical field of data mining, and in particular relates to a classification method based on multi-view dictionary learning for mixed sampling industrial big data. Background technique [0002] With the rapid development of modern industry, the production equipment in modern enterprises is becoming larger, continuous and automated, the structure or composition of equipment is becoming more and more complex, and the collection, source and form of production data are becoming more and more diverse. In the actual production, when obtaining the electrolytic cell data of industrial aluminum production, the frequency of data collection will be different due to the cost of collection. For example, the data of feature sets such as iron content, silicon content, molecular ratio, and electrolyte level require experts to collect experimental data, which is expensive, expensive, and low in sampling frequency; while the data of feature...

AI Technical Summary

Problems solved by technology

Therefore, the present invention intends to learn from the idea or method of processing multi-view data to solve the classification problem in mixed sampling industrial big data, so as to overcome the inconsistency of data collection frequency caused by factors such as high data collection cost and high cost in industrial big data

Images