Abstract Views :204 |
PDF Views:0
Authors
Source
Data Mining and Knowledge Engineering, Vol 10, No 8 (2018), Pagination: 160-166
Abstract
The widespread of social media websites resulted to produce a massive amount of data every single minute. This kind of data represents people’s opinions, attitudes and feedback about different topics, political decisions, and products. Processing and analyzing such kind of data in order to understand people’s thoughts, feedback, and needs is called sentiment analysis (opinion mining). Sentiment analysis becomes a hot area nowadays because of this rapid growth in social media websites. Sentiment analysis main task is classifying text/documents/words according to its polarity into positive/negative opinions. The traditional sentiment analysis techniques are supervised, unsupervised and semi-supervised. Theses traditional techniques still didn’t give a high quality while working with Arabic language. Due to the complexity of Arabic language and its high derivatives, the sentiment analysis task becomes complicated and not an easy task. This paper applies different supervised techniques (machine learning algorithms) with different preprocessing methods in order to investigate their importance.
Keywords
Sentiment Analysis, Preprocessing, Stemming, Arabic Sentiment Analysis, NLP, Features.
