Classification of emotions induced by music videos and correlation with participants’ rating

作者：

Highlights：

• The DT-CWPT time–frequency features represent user’s emotional state effectively.

• SVD–QRcp and F-Ratio based feature selection eliminate redundant and weak features.

• Features selected are mostly from brain region involved in emotional activities.

• Valance and liking show strong correlation in almost all subbands of DT-CWPT.

摘要

•The DT-CWPT time–frequency features represent user’s emotional state effectively.•SVD–QRcp and F-Ratio based feature selection eliminate redundant and weak features.•Features selected are mostly from brain region involved in emotional activities.•Valance and liking show strong correlation in almost all subbands of DT-CWPT.

论文关键词：Dual-Tree Complex Wavelet Packet Transform (DT-CWPT),Electroencephalogram (EEG),Singular Value Decomposition (SVD),QR factorization with column pivoting (QRcp),F-Ratio,Support Vector Machine (SVM),Human-computer interaction (HCI)

论文评审过程：Available online 12 April 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.03.050