Method and system for classifying machine error data

Abstract

The present application relates to the technical field of data mining, in particular to a method and system for classifying machine error data. This method introduces the label propagation algorithm as a preprocessing step of machine error data, uses a small number of calibrated machine data labels, quickly estimates the labels of uncalibrated data, and forms a classification training set. Based on the machine data and labels in the classification training set, the dictionary learning with consistent labels is carried out, and at the same time, the reconstruction error, sparse coding error and classification error are minimized, and the reconstruction dictionary, sparse coding and multi-class linear classifier are obtained, while maintaining the dictionary. The relationship between items and data labels. Furthermore, the obtained sparse coding is used to represent the characteristics of machine data, and input to the classifier for prediction, to determine the category of test samples, and to realize the misclassification of machine data. By introducing efficient semi-supervised data preprocessing, the number of calibrated machine samples is increased, the prior information is enriched, and the accuracy of machine data classification is effectively improved.

Description

technical field [0001] The present application relates to the technical field of data mining, in particular to a method and system for classifying machine error data. Background technique [0002] With the continuous development of computer technology and intelligence, machine error data classification has become a very important research topic in data mining. Among them, the machine error data classification technology describes the machine error data, analyzes the data structure, and then obtains the data characteristics, and finally classifies the machine error data according to the data characteristics. [0003] In the prior art, dictionary learning algorithms, such as K-SVD and D-KSVD (Discriminative K-SVD), are usually used to learn the machine data in the training set and the calibration of the machine data to obtain a reconstructed dictionary, sparse coding And the classifier, and use the obtained sparse coding to represent the characteristics of the machine data, a...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the present application provides a machine error data classification method and system to solve the problem that when the number of samples in the selected training set is small in the prior art, it is easy to cause the characteristics of the machine error data to be accurately represented, and to reduce the machine error data. The problem of classification accuracy

Images