An adaptive model for knowledge mining in databases “EMO_MINE” for tweets emotions classification

Abstract

knowledge Mining from large databases has been recognized as a key research topic in database systems and machine learning, many researchers take in their consideration the importance of knowledge extraction from the useful databases. Twitter nowadays becoming a magnificent wide space for getting people’s opinions, sentiments and emotions, manufactures aims to knowing their costumers opinions about a specific product, since millions of people connecting through social media pages like twitter for every day and sharing their opinions and emotions towards a product or an event, but there are no enough experiments go further to investigate and characterize the feelings behind tweets. This paper aims to exploring how to facilitate extracting emotions from text tweets by presenting an adaptive model for extracting and classifying emotions in Arabic tweets (EMO_MINE), based on four emotions sad, joy, happy and anger. The unique value of this model is based on the integration of SQL and machine learning techniques. The experimental results demonstrate how this proposed model for extracting emotions and opinions is useful for knowledge discovery of Arabic tweets using SQL and machine learning algorithms. The results show that our proposed model is improve the classification process in term of accuracy and ROC measurements, the naïve base classifier gives a very satisfy results comparing with others classifiers that’s examined in this study.