Power equipment defect text feature classification method based on combinatorial algorithm

Abstract

The invention discloses a power equipment defect text feature classification method based on a combinatorial algorithm. The method comprises the following steps: establishing a dictionary and a stop word table according to defect description of power equipment; performing Jieba word segmentation on the defect text; representing the text subjected to word segmentation in a distributed manner according to a word vector representation method; carrying out weight distribution according to an Attention mechanism, carrying out further extraction of features according to a CNN algorithm, and then carrying out feature weight distribution according to the Attention mechanism to obtain Attention-CNN extraction features; extracting text features according to an RNN algorithm to obtain RNN extraction features; and fusing and inputting the two extracted features into an SVM classifier to obtain the division of the defect features of the power equipment. According to the method, the text features are subjected to weight distribution, the two extracted features are fused, comprehensive extraction of the text features and distribution of feature weights are achieved, and therefore the accuracy of power equipment defect text classification is effectively improved.

Description

technical field [0001] The embodiment of the present application relates to the field of classification of defect texts of electric equipment, and in particular to a method for classifying defect text features of electric equipment based on a combined algorithm. Background technique [0002] During the long-term operation of power equipment, a large amount of defect text data will be generated. These text data are very important to the evaluation of the operating status of power equipment, but because the text data is difficult to process, a large number of power equipment defect classification needs to be done manually, which not only limits the work efficiency, but also due to the differences in the knowledge level and experience of the staff, It is difficult to accurately judge the defect level of electrical equipment. At present, according to the defect description text, the defect characteristics of power equipment are divided into three levels: general, major, and urg...

AI Technical Summary

Problems solved by technology

[0005] This application provides a method for classifying text features of electrical equipment defects based on combined algorithms to solve the problems of weak text feature expression and low classification accuracy of traditional algorithm models in the prior art

Images