LSTM Based Emotion Analysis of Text in Tamil Language

Abstract

The sentiments and emotions expressed by users on the internet greatly influence the decision-making process of business firms. Recent studies show that emotion analysis yields more precise information than sentiment analysis. Text emotion analysis has become popular for higher-demand languages like English, Chinese, French, and Arabic. However, no prior studies have been conducted on locally speaking languages, including Tamil, Malayalam, and Sinhala. Therefore, this paper presents a deep learning based novel model to identify the emotions expressed in Tamil texts using a Long Short-Term Memory (LSTM) network. Besides, to enhance the robustness of our proposed model, we conducted experiments with machine learning classifiers, including Support Vector Machine (SVM), Naïve Bayes (NB), Logistic Regression (LR), and Random Forest Classifier (RFC). The experimental results prove that our Tamil text emotion analysis model significantly outperforms other machine learning models, achieving an accuracy of 80%.