基于TF-IDF和FastText的快速外卖评论情感分类研究

Study on the Fast Emotional Classification of Takeout Reviews Based on TF-IDF and FastText

对外卖评论进行情感分类在外卖评论挖掘、兴趣推荐等领域都有很高的价值,但外卖数据参差不齐,过长的外卖评论会导致模型难以提取准确文本特征等问题.因此设计了一种基于TF-IDF和FastText的外卖评论情感分类模型并进行了相关实验,实验结果表明,该模型在3s内达到了90.23%的准确率,具有训练速度快,准确率高的特点,能够快速的对外卖评论进行情感分类.

Emotional classification of takeout comments has high value in takeout comment mining,interest recommendation and other fields,but takeout comments are different in length.Too long takeout comments will make it difficult for Fasttext model to extract accurate text features.Therefore,this paper discusses a sentiment classification model of takeout comments based on TF-IDF and FastText and conducts related experiments.The experimental results show that this model achieves 90.23%accuracy in 3s,has the characteristics of fast training speed and high accuracy,and can quickly classify the emotion of takeout comments.