Sentiment Learning on Product Reviews via Sentiment Ontology Tree

摘要: Existing works on sentiment analysis product reviews suffer from the following limitations: (1) The knowledge of hierarchical relationships products attributes is not fully utilized. (2) Reviews or sentences mentioning several associated with complicated sentiments are dealt very well. In this paper, we propose a novel HL-SOT approach to labeling product's and their in by Hierarchical Learning (HL) process defined Sentiment Ontology Tree (SOT). empirical against human-labeled data set demonstrates promising reasonable performance proposed approach. While paper mainly one product, our easily generalized mix more than products.