National Research Council of Canada. Information and Communication Technologies
Text, Book Chapter
16th International Conference on Intelligent Text Processing and Computational Linguistics (CICLing 2015), April 14-20, 2015, Cairo, Egypt
linguistics; text processing; co-training; co-training algorithm; product reviews; computational linguistics
Opinions expressed about a particular subject are often nuanced: a person may have both negative and positive opinions aboutdifferent aspects of the subject of interest, and these aspect-specific opinionscan be independent of the overall opinion. Being able to identify,collect, and count these nuanced opinions in a large set of data offersmore insight into the strengths and weaknesses of competing productsand services than does aggregating overall ratings. We contribute a newconfidence-based co-training algorithm that can identify product aspectsand sentiments expressed about such aspects. Our algorithm offers betterprecision than existing methods, and handles previously unseen languagewell. We show competitive results on a set of opinionated sentences aboutlaptops and restaurants from a SemEval-2014 Task 4 challenge.
Springer International Publishing
Computational Linguistics and Intelligent Text Processing: 16th International Conference, CICLing 2015, Cairo, Egypt, April 14-20, 2015, Proceedings, Part II9042: 225–240.