Research on Barrage Sentiment Ontology Construction Based on SO-PMI Algorithm

Barrage is the commentary content that pops up when a video playing, which is short and concise and has strong domain characteristics. Because the sentiment analysis method in the field of barrage language is relatively single, this paper presents a method for constructing a library of barrage ontology based on SO-PMI algorithm and the word vector maximum similarity algorithm. The barrage ontology is based on professional knowledge and the barrage sentiment dictionary, which is added into the ontology as an example to construct the knowledge base. The sentiment classification of barrage can be realized and a new way for sentiment analysis of barrage language is provided by this method. The barrage ontology with tendency of sentiment can be better applied to video recommendation and opinion monitoring.