TY - GEN
T1 - Unsupervised subjectivity-lexicon generation based on vector space model for multi-dimensional opinion analysis in blogosphere
AU - Chen, Hsieh Wei
AU - Lee, Kuan Rong
AU - Huang, Hsun Hui
AU - Kuo, Yaw Huang
PY - 2010
Y1 - 2010
N2 - This paper presents an unsupervised framework to generate a vector-space-modeled subjectivity-lexicon for multi-dimensional opinion mining and sentiment analysis, such as criticism analysis, for which the traditional polarity analysis alone is not adequate. The framework consists of four major steps: first, creating a dataset by crawling blog posts of fiction reviews; secondly, creating a "subjectivity-term to object" matrix, with each subjectivity-term being modeled as a dimension of a vector space; thirdly, feature-transforming each subjectivity-term into the new feature-space to create the final multi-dimensional subjectivity-lexicon (MDSL); and fourthly, using the generated MDSL for opinion analysis. In the experiments, it shows that the improvement by the feature transform can be up to 31% in terms of the entropy of features. In addition, the subjectivity-terms and objects are also successfully and reasonably clustered in the demonstration of fiction review (literary criticism) analysis.
AB - This paper presents an unsupervised framework to generate a vector-space-modeled subjectivity-lexicon for multi-dimensional opinion mining and sentiment analysis, such as criticism analysis, for which the traditional polarity analysis alone is not adequate. The framework consists of four major steps: first, creating a dataset by crawling blog posts of fiction reviews; secondly, creating a "subjectivity-term to object" matrix, with each subjectivity-term being modeled as a dimension of a vector space; thirdly, feature-transforming each subjectivity-term into the new feature-space to create the final multi-dimensional subjectivity-lexicon (MDSL); and fourthly, using the generated MDSL for opinion analysis. In the experiments, it shows that the improvement by the feature transform can be up to 31% in terms of the entropy of features. In addition, the subjectivity-terms and objects are also successfully and reasonably clustered in the demonstration of fiction review (literary criticism) analysis.
UR - https://www.scopus.com/pages/publications/77958502594
UR - https://www.scopus.com/pages/publications/77958502594#tab=citedBy
U2 - 10.1007/978-3-642-14922-1_46
DO - 10.1007/978-3-642-14922-1_46
M3 - Conference contribution
AN - SCOPUS:77958502594
SN - 3642149219
SN - 9783642149214
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 372
EP - 379
BT - Advanced Intelligent Computing Theories and Applications - 6th International Conference on Intelligent Computing, ICIC 2010, Proceedings
T2 - 6th International Conference on Intelligent Computing, ICIC 2010
Y2 - 18 August 2010 through 21 August 2010
ER -