Sentiment Analysis through Machine Learning: An Experimental Evaluation for Albanian

Opinions have always influenced our behaviours and they have a key role in human activities. Nowadays, online opinion resources such as newspapers, blogs, and reviews have enormously increased the amount of text data available for analysis. Sentiment Analysis (or Opinion Mining) is increasingly becoming an important tool for analysing text data in order to understand opinions correctly. In this context, machine learning methods have the potential to perform correct classification of texts as expressing positive or negative opinion for a certain topic. However, much research has been dedicated to languages such as English, Japanese, Chinese or German but no research has been made for other rare Indo-European languages such as the Albanian. In this paper, we present the first approach for Sentiment Analysis in Albanian. We show through extensive experiments with text data from political news consisting of five different topics, that the proposed approach is effective in classifying text documents as belonging to negative or positive opinion regarding the given topic.