摘要
情绪一直受到研究者的广泛关注,但是关于情绪自动识别的研究却很少。通过收集自由文本输入时的击键特征、停顿特征和修改特征,使用分类算法判断输入者是否出现喜、怒、哀、惧四种基本情绪,可以建立情绪的自动识别模型。实验结果表明,四种基本情绪都能通过文本输入的特征进行推断,并且拥有较高的准确率。
Emotion is widely researched, but little has been done on automatic emotion recognition. In this paper, we propose an approach that in- tends to recognize emotion automatically, we collected keystroking characteristics, pausing features and modifying features, and used clas- sification algorithms to determine whether the user expressed the four kinds of basic emotions, happiness, anger, sadness and fear. Result showed that the four basic emotions can be inferred through the characteristics, and it performed a high accuracy rate.
出处
《现代计算机》
2017年第3期3-6,11,共5页
Modern Computer
关键词
情绪
自由文本
击键特征
分类算法
自动识别
Emotion
Free Text
Keystroke
Classification Algorithm
Automatic Identification
作者简介
曹晓春（1991-），男，四川广安人，硕士研究生，研究方向为人机交互
