Big Data, Sentiment Analysis, Machine learning, Classification

October, 2017

The evolution of big data shaped innumerable possibilities in data analysis. Sentiment analysis (SA) refers to opinion mining which main task is to categorize public opinion into likely and unlikely opinions generally referred as positive and negative opinions. Another task of SA is to found the subjective and the objective expression of the writer’s opinion about the given text. The sentiment analysis can be performed at different levels such as document, sentence and word levels. This study focuses on the different technologies which are currently used in SA such as machine learning techniques including supervised, unsupervised and semi-supervised learning techniques, lexicon based and hybrid techniques are most recently used in current research. We found that recent techniques are still unable to work on different languages and domains; current techniques are still unable to deal with complex sentences; classification based on labelled data set is also a challenging problem and requires more effort in this field.

32-37

.-- (Online) - -- (Print)

Vol.1 - No.1

Click here to download full paper

http://dx.doi.org/