通常,在我看到所使用的“影响分类”一词时,它描述语音数据的处理,而不是使用像频谱、大声等音频特征来确定某一发言者是否愤怒、沮丧等等。 这些特点显然不会很好地翻译成文。
When speaking of classifying text, its more common to refer to classifying Sentiment, aka "Opinion Mining" to determine if the author of a text is speaking positively or negatively about the text s subject. If this provides enough nuance for you needs, fortunately there are a large amount of resources to help you with this. In Python, the Natural Language Toolkit provides classifiers that are often used for this type of work, for example this demo.
这种做法的倒数部分是,它一般限于积极的/消极的分类,而且非常具体。 例如,在软件审查分类方面,受过检测正面电影审查培训的班级人员将表现不佳。
http://www.google.com/url?sa=t&rct=j&q=feeler%20textctor%20classification&=web&cd=1&sqi=2&ved=0CQFjA&url=%3A%2Fciteseerx.psu.edu2 不幸的是,你似乎不太可能发现有太多的图书馆对这项任务的支持,但鉴于研究中的信息,你应当能够在以下几个方面建立这种系统:NLTK或其他分类/天然语言处理图书馆。