An electrocardiosignal classification method based on optimal feature space similarity mining label correlation

Abstract

The invention discloses an electrocardiosignal classification method based on optimal feature space similarity mining label correlation, and belongs to the field of intelligent diagnosis of electrocardio diseases. According to the invention, on the basis of a trained single label classifier, label correlation mined by utilizing an optimal feature space similarity algorithm to correct association rules is proposed, and a final prediction label is determined according to the corrected label correlation; the method comprises the following specific steps: 1, obtaining a main label and a candidateauxiliary label set by utilizing a classification result of a classifier; and 2, mining label correlation by adopting an association rule, correcting the label correlation by utilizing an optimal feature space similarity algorithm provided by the invention, and filtering the candidate sub-label set through the corrected label correlation to determine a final prediction label. According to the invention, the relation between the labels is corrected by calculating the optimal feature space similarity between the different labels, more accurate label correlation is obtained, and the precision ofelectrocardiosignal classification is improved.

Description

technical field [0001] The invention belongs to the field of intelligent diagnosis of electrocardiographic disorders, and in particular relates to a machine learning-based multi-label disease determination method, in particular to a method for classifying electrocardiographic signals based on optimal feature space similarity mining for correlation between tags. Background technique [0002] In recent years, multi-label learning has gradually become one of the hot research issues in the field of machine learning. Unlike traditional single-label classification where each sample belongs to only one class label, each sample in multi-label classification belongs to multiple class labels. The multi-label problem is defined as: Let X=R d Represents the d-dimensional sample space, L={l 1 , l 2 ,... l n} represents a label space containing n labels, D={(x i ,Y i )|1≤i≤m,x i ∈X,Y i ∈L} represents a training set containing m samples, where x i =[x i1 ,x i2 ,...,x id ] repre...

AI Technical Summary

Problems solved by technology

[0006] Algorithm adaptation strategies and problem conversion strategies largely ignore the correlation between labels, and do not use the relationship between labels to build models, and there is such a relationship between ECG disorders, so these methods cannot be used very well. Using ECG to determine the condition, the accuracy of prediction is poor

Images