Speech emotion classification method based on feature fusion and ensemble learning

• Main Authors: HUANG YONGMAO, HUANG WENTAO, XU LIANG, QING CHAOJIN, GUO YI, XIONG XUEJUN
• Format: Patent
• Language: Chinese; English
• Published: 28.05.2021
• Subjects: ACOUSTICS; CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; MUSICAL INSTRUMENTS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION

The invention discloses a speech emotion classification method based on feature fusion and ensemble learning, and the method comprises the following steps: collecting a plurality of speech data, and carrying out the preprocessing of the speech data; performing feature extraction on the preprocessed data, and constructing a feature set; adopting multiple classifiers to construct an integrated learning classification model, and training the integrated learning classification model; and adopting the trained integrated learning classification model to identify the feature set corresponding to the to-be-identified voice data, obtaining a classification result, and obtaining a voice emotion classification result. According to the method, the voice emotion of the speaker can be effectively predicted and classified through the voice data. 本发明公开了一种基于特征融合与集成学习的语音情感分类方法，包括以下步骤：采集若干语音数据，并对语音数据进行预处理；对预处理后的数据进行特征提取，并构建特征集；采用多分类器构建集成学习分类模型，并对集成学习分类模型进行训练；采用训练后的集成学习分类模型对待识别语音数据对应特征集进行识别，获取分类结果，得到语音情感分类结果。本发明能够有效地通过语音数据对说话人的语