基于文本挖掘的商品推荐

摘要

对电子商务网站的评论文本进行分词、去停用词等整理,通过词频统计提取特征词,应用词频-逆文档频率提高特征词的类别区分能力以增加特征词的准确性.在收集大量的电子商务网站的评论文本及一系列预处理后构建了特征词词库.采用词语相似度计算方法用于关键词向量与特征词词库相似度的计算.根据相似度计算结果对用户评论的商品进行排序以实现对用户商品的推荐.设计了商品推荐系统并完成了实验程序.利用收集到的用户评论文本完成了对商品的推荐实验,并对实验结果进行了考察与分析.%By dealt with the user comment text of the electronic commerce website, such as word segmentation, removing stop words and word frequency statistics, the character-words were extracted.Then TF-IDF was applied to improve word feature category distinguishing ability.After a large number of electronic commerce comment text of website were collected, characteristic word corpus was constructed.The word similarity calculation based thesaurus was used for similarity calculation of keyword vector and character-words.According to the results of the similarity calculation of the user comments with characteristic word, rank of the purchase products and recommendation were carried out.Finally, the commodity recommendation system was designed and implemented, the experiment was carried out and the experimental results were investigated and analyzed.