Application Research On Semantic Classification In Taobao Comment Text

The rapid development of electronic commerce in recent years, more and more people choose online shopping platform, taobao, as the largest C2 C e-commerce platform with a large customer base, people can buy goods for thousands of miles away through taobao in a short period of time without having to field. This affected the buying behavior of people to a certain exten, people can communicate through the network to shorten the distance between people, which makes people search the commodities or services related to assessment information before buying behavior habits,, the value of consumer online reviews text were dug up, In addition merchants can also improve to ascend the commodities or services obtained from the comments on the commodity information,then further to extract the effective information of user evaluation is the primary issue.In the present study, different scholars have various angles were carried out for the emotion classification research, emotional words and negative features, feature level field collection of emotional mining and a series of viewpoints and methods have been put forward. But for the characteristics of the comment text, traditional text classification technology is not suitable for this. For the comment text classification, emotional tendency classification and so on were included in the study and put forward a series of applied research.So this article puts forward the comment text classification method based on semantics. The emotional tendencies of mining the comment text words, and classificate through the emotional concept see classification lexical semantic relations. On the basis of the original emotion classification was improved, using "hownet" and "a synonym for Lin" build the initial relationship between vocabulary of an ontology. And combined with the emoticons, dictionary and network language emotion, along with the development of the words added ontology corpus. Determine the emotion tendentiousness of seed words, again through the other and seed words relations between concepts to determine other vocabulary of emotional tendency.Semantic praise and tendency of research provides a new train of thought and method to text classification, text filtering and other natural language processing research, the applications of semantic classification to taobao in the comment text will also is a kind of trend. Through semantic classification standard will be reclassified the comment text. Change the original manual selection "praise", "medium review", "bad review" to automatically identify "or" "in" "fall" "did not evaluate the classification of the standard. The purpose of this paper is to pass the test text for empirical research, verify the feasibility of semantic classification. In empirical research of the purpose of this paper is to pass the test text, by means of semantic classification after the text is more intuitive, clear, so as to validate the feasibility of semantic classification.