Sentiment Analysis Using Sentiment Features
A novel solution to SA of short informal texts with a main focus on Twitter posts known as "tweets" is introduced, and it is illustrated that the hybrid method outperforms the state-of-the-art unigram baseline.
Abstract
Sentiment Analysis (SA) or opinion mining has recently become the focus of many researchers, because analysis of online text is beneficial and demanded for market research, scientific surveys from psychological and sociological perspective, political polls, business intelligence, enhancement of online shopping infrastructures, etc. This paper introduces a novel solution to SA of short informal texts with a main focus on Twitter posts known as "tweets". We compare state-of-the-art SA methods against a novel hybrid method. The hybrid method utilizes a Sentiment Lexicon to generate a new set of features to train a linear Support Vector Machine (SVM) classifier. We further illustrate that our hybrid method outperforms the state-of-the-art unigram baseline.