Application Of Text Mining In The Analysis Of Online Comments

With the development of online shopping, there is more and more online shopping comments.However, the initial comment on the online shopping page is in the form of a sentence, so, it's difficult to analyze.It's more difficult to intuitively find valuable information in content.Therefore, how to use these huge amount of comments to extract valuable information for businesses to improve their service, to enhance their sales and profits is essential.This paper solves this problem from the perspective of text mining.First of all, the data used in this paper is extracted by web crawlers from MIDEA water heater's product comments,the extracted comments are processed to get the data matrix.,so,it's very valuable to get the first-hand data.Secondly,an emotional score algorithm is designed in this paper,which uses the method of key words and the word order to get a good result. This result is consistent with the actual emotion of the comment. thirdly, this paper uses text mining methods, such as decision trees,logistic regression and association rule. These methods are very effective in the analysis of the comments by important words.Finally, this paper put forward the strategy of improving the goods and profit,so the conclusion of this paper has realistic meaning.The conclusion of this paper is the analysis of the concerns of customers and what is the advantages of goods and what is the disadvantages by using text mining methods.These analyses are valuable and innovative research.At the end of this paper,the results of this analysis are summarized.