Based on SO-PMI Algorithm to Discriminate Sentimental Words´ Polarity in TV Programs´ Subjective Evaluation

With the increasing of comment data about television programs on Internet, using text mining technology to analyze the massive data so that it can make subjective evaluation about television programs. It can give decision support and improve competitiveness for each department in radio and TV industry. This paper mainly introduces how to using sentiment-oriented pointwise mutual information (SO-PMI) to judge sentimental words´ polarity and calculate sentimental words´ intensity. It lays a solid foundation for the further study of subjective evaluation about television programs.