中文社会媒体环境下半监督学习的汽车缺陷识别方法

Semi-supervised Learning for Automobile Defect Identification in the Context of Chinese Social Media

随着国内汽车产业的不断发展以及社会媒体的普及,从社会媒体中获取用户反映的汽车缺陷的重要性日益凸显。本文针对现有汽车缺陷发现研究中存在的分类过程需要大量人工标注,以及根据产品部件对缺陷内容进行分类不完全适用等问题,构建了基于中文社会媒体的汽车缺陷识别框架及汽车缺陷特征集,研究了利用Tritraining半监督学习算法进行汽车产品缺陷分类的方法以及利用LDA进行汽车产品缺陷主题建模的方法,并进行了相关实验。实验结果表明,本文所提出的方法能够有效地识别出中文社会媒体中的汽车缺陷,并在效率和精准度上都有较好的表现。

With the continuous development of the domestic auto industry and the popularity of social media,the importance of user feedback on the car defects from social media become increasingly prominent.For existingissues in recent product defect discovering research,such as the classification process requires a lot of manual annotation and classification according to product content is not fully applicable.The framework for automobile defect identification and the feature set of automobile defect are constructed in the context of Chinese social media.Furthermore,tri-training semi-supervised learning algorithm-based automotive product defect classification methods and the use of automotive product defects LDA topic modeling method are researched in this paper.Experimental results show that the proposed method can effectively identify the automobile defect in the context of Chinese social media,and has better performance in terms of both efficiency and accuracy.