基于LDA 模型与语义网络对评论文本进行挖掘

Research on Mining Comment Text Based on LDA Model and Semantic Network

本文基于LDA主题聚类与语义网络模型LTC-SNM的方法,依据客户价值理论对时尚租赁APP在线评论文本进行研究;对在线评价文本进行数据清洗和关键特征提取,使用Word2vec生成词向量,利用分类器对评价文本进行情感二分类,再通过LDA模型进行文本聚类,提取在线文本主题词转换为共线矩阵,最后利用ROSTCM软件生成语义网络图。实验结果表明:基于LTC-SNM的挖掘方法更能全方位地表达在线用户评论的主题。

Based on customer value theory and LDA subject clustering and semantic network model LTC-SNM,this paper studies the online review text of clothing rental APP.The online evaluation text was obtained for data cleaning and key feature extraction.Word2vec was used to generate word vectors.The classifier was used for emotional dichotomization of the evaluation text.ROSTCM software is used to generate semantic network graph.Experimental results show that the mining method based on LTC-SNM can more comprehensively express the topic of online user comments.