Emotion feature optimisation based on PCA-GRA analysis
by Guohua Hu; Guoyan Meng; Qingshan Zhao; Xiaoxia Zheng
International Journal of Computing Science and Mathematics (IJCSM), Vol. 12, No. 4, 2020

Abstract: The interference and redundancy of speech emotional features will directly affect the recognition performance of emotional features. In order to enhance the ability of emotional features to recognise speech emotion, dimension reduction method was used to optimise emotional features. Four emotion (sad, angry, happy and neutral) voices were selected from Berlin speech database and CASIA corpus, and traditional emotion features (prosodic features, formant features and mel frequency cepstrum coefficient (MFCC) features) were extracted. In order to reduce the mutual interference between features, principal component analysis (PCA) was used to extract the principal components of features to get independent features. Meanwhile, in order to get the emotional features that are highly related to the emotional type, grey relational analysis (GRA) was used to select the main features from the principal components and design experiments for comparison. The experimental results show that PCA-GRA dimensionality reduction method can reduce the correlation of features and get the feature set with small redundancy, so as to improve the recognition effect of speech emotion.

Online publication date: Tue, 26-Jan-2021

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Computing Science and Mathematics (IJCSM):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com