Prediction of Emoji from News Headlines using Machine Learning Techniques

Abstract

Emojis are generated from the news headlines of Malayalam language using Naive Bayes (NB) and Support Vector Machine(SVM) classifiers. The human brain processes visual data faster than text data. Assigning emoji to the text helps the people easily categorize the news based on the emotion without reading the entire sentence. Six different emojis are assigned to the news headlines based on the emotional contents of the text. These emojis are used for representing emotions like sad, angry, fear, happy, love, and neutral. The dataset contains 3111 sentences which are retrieved from the tweets of Manorama Online. Both Bag-of-Words (BOW) and Term Frequency versus Inverse Document Frequency (TFIDF) features are used for feature vector formation of the dataset. The SVM shows better accuracy compared with the NB classifier.