This paper reports on the use of ensemble learning to classify the sentiment of tweets as being either positive or negative. Tweets were chosen because Twitter is both a popular tool and a public, human annotated dataset was made available as part of the SEMVAL 2013 competition. We report on an approach to classification thatcontrasts single machine learning algorithms with a combination of algorithms in an ensemble learningapproach. The single machines learning algorithms used were Support Vector Machine (SVM) and Naïve Bayes (NB) while the method of ensemble learning was the arbiter tree. Our system achieved an F score using thearbiter tree at 83.55% which was the same as SVM but quite slightly than Naïve Bayes algorithm.
|Number of pages
|International Journal of Advances in Engineering Technology
|Published - Jan 2015