E2SAM: Evolutionary ensemble of sentiment analysis methods for domain adaptation

摘要: Abstract Currently, a plethora of industrial and academic sentiment analysis methods for classifying the opinion polarity text are available ready to use. However, each those have their strengths weaknesses, due mainly approach followed in design (supervised/unsupervised) or domain used development. The weaknesses usually related capacity generalisation machine learning algorithms, lexical coverage linguistic resources. Those issues two main causes one challenges Sentiment Analysis, namely adaptation problem. We argue that right ensemble set heterogeneous Analysis Methods will lessen Thus, we propose new methodology optimising contribution off-the-shelf an classifier depending on input text. results clearly show our claim holds.