Neural structure corresponding learning cross-domain emotion classification method for improving feature selection

Abstract

The invention relates to a neural structure corresponding learning cross-domain emotion classification method for improving feature selection, and belongs to the field of natural language processing.The method comprises the following steps: firstly, selecting two different fields in an Amazon comment data set as a source domain and a target domain, preprocessing source domain and target domain data to obtain text contents of the source domain and the target domain, secondly, carrying out word form restoration on a text, eliminating redundant features, and carrying out vectorization processingon the text to obtain initial features of the text; screening out pivot features through a chi-square test feature selection method to serve as pivot features in cross-domain tasks, and the rest features being non-pivot features; performing pivot feature prediction on the non-pivot features in the two fields through neural structure corresponding learning by utilizing the obtained pivot featuresto obtain feature migration; and training a logistic classifier by using the initial features and the migration features of the source domain text, and testing by using the text features and the migration features of the target domain to obtain a classification result of the target domain.

Description

technical field [0001] The invention relates to a neural structure correspondence learning cross-domain emotion classification method for improved feature selection, and belongs to the technical field of natural language processing. Background technique [0002] A domain refers to a class of entities, and different domains are different types of entities. For example, BOOK and DVD can be regarded as different domains. Because product reviews have strong domain characteristics. The emotional information expressed by the same word in the two domains is not necessarily the same. Additionally, sentiment classifiers trained on labeled corpora in one domain have reduced classification accuracy in another. In the domain adaptation work of feature transfer, the most typical method is Structural Correspondence Learning (SCL), which establishes the correlation between the source domain and the target domain through pivot features based on different domains. However, when the tradi...

AI Technical Summary

Problems solved by technology

However, when the traditional method calculates the pivot feature, the method of screening the pivot feature only uses the mutual information value MI of the feature. This method only considers the degree of association between the pivot feature and the classification label, and does not consider the pivot feature for the text. The degree of importance and feature redundancy, so the selected pivot features are not reasonable enough, which affects the feature migration results

Images