Title :
Optimizing N-gram based text feature selection in sentiment analysis for commercial products in Twitter through polarity lexicons
Author :
Cabanlit, Mark Anthony ; Junshean Espinosa, Kurt
Author_Institution :
Dept. of Comput. Sci., Univ. of the Philippines Cebu, Cebu, Philippines
Abstract :
This study aims to optimize N-gram based text feature selection in sentiment analysis for commercial products in twitter through polarity lexicons. This can be done by merging dictionary-based weighing with naïve-Bayes classification of sentiments. The study is still ongoing but partial results show potential.
Keywords :
information analysis; learning (artificial intelligence); pattern classification; social networking (online); N-gram based text feature selection; Twitter; commercial products; dictionary-based weighing; naive-Bayes classification; polarity lexicons; sentiment analysis; Computer science; Dictionaries; Market research; Media; Sentiment analysis; Twitter; Bigram; N-gram; Naïve Bayes; Opinion Classification; Sentiment Analysis; Social Media; Trigram; Twitter; Unigram;
Conference_Titel :
Information, Intelligence, Systems and Applications, IISA 2014, The 5th International Conference on
Conference_Location :
Chania
DOI :
10.1109/IISA.2014.6878767
