Language Independent Word Polarity Detection using Star Rating

The continuous and rapid evolution of the Internet along with increased accessibility to social media platforms has led to an exponential growth in the amount of content posted online by users daily. Sentiment Analysis (SA), a field of Natural Language Processing (NLP), is nowadays necessitated to be performed on user-generated content in order to extract valuable information. This paper proposes a methodology to determine the polarity of a word based on its occurrence in positive / negative product reviews. The star rating of a review is used to classify it as positive or negative. The proposed methodology has been implemented using a German language Amazon product review dataset, although the algorithm is independent of any language. F1-score thus obtained is 72% which is better than the lexicon-based baseline sentiment analysis for German language