基于机器学习的语音情感识别
首发时间:2020-01-02
摘要:随着计算机技术的发展和人工智能的普及,语音情感识别研究收到学界和工业届的广泛关注。从语音情感识别的起源、语音情感的分类,到研究现状进行归纳与总结。目前的情感识别任务大多采用人工提取多种声学特征并物理降维,构建特征工程的方法,提升识别结果,于是对机器学习常用的几种算法进行调研,例如:支持向量机、决策树、支持向量机等。调研这些机器学习算法从原理到应用,并进行详细阐述。使用其中几种机器学习算法训练语音情感识别模型,利用模型进行语音情感分类。根据实验结果对比各个算法的性能,选择出其中最适合语音情感识别的机器学习算法。
For information in English, please click here
Speech Emotion Recognition Based on Machine Learning
Abstract:With the development of computer technology and the popularization of artificial intelligence, the study of speech emotion recognition has received extensive attention from academics and industry. From the origin of speech emotion recognition, the classification of speech emotions, to the research status, it is summarized and summarized. Most of the current emotion recognition tasks use manual extraction of multiple acoustic features and physical dimensionality reduction, constructing feature engineering methods to improve recognition results, and then investigating several algorithms commonly used in machine learning, such as support vector machines, decision trees, and support Vector machine and so on. Investigate theseSpeech Emotion Recognition Based on Machine Learning machine learning algorithms from principle to application and elaborate. Use several of these machine learning algorithms to train speech emotion recognition models and use the models for speech emotion classification. According to the experimental results, the performance of each algorithm is compared, and the machine learning algorithm that is most suitable for speech emotion recognition is selected.
Keywords: Speech Emotion Recognition Discrete Emotion Machine Learning Support Vector Machine
基金:
引用
No.****
动态公开评议
共计0人参与
勘误表
基于机器学习的语音情感识别
评论
全部评论0/1000