Optimization of Sentiment Analysis Using Teaching-Learning Based Algorithm

Feature selection and sentiment analysis are two common studies that currently being conducted; consistent with the advancements in computing growing use of social media. High dimensional or large feature sets is a key issue as it can decrease accuracy classification make difficult to obtain optimal subset features. Furthermore, most reviews from media carry lot noise irrelevant information. Therefore, this study proposes new text-feature method uses combination rough set theory (RST) teaching-learning based optimization (TLBO), which known RSTLBO. The framework develop proposed RSTLBO includes numerous stages: (1) acquiring standard datasets (user six major U.S. airlines) used validate search result methods, (2) pre-processing dataset using text processing methods. This involves applying methods natural language techniques, combined linguistic techniques produce high results, (3) employing method, (4) selected features previous process for Support Vector Machine (SVM) technique. Results show an improvement when combining processing. More importantly, algorithm able improved analysis.