Deep Learning Technique for Sentiment Analysis of Hindi-English Code-Mixed Text using Late Fusion of Character and Word Features

Enormous presence of Code-Mixed text in social media provides an important research area for Natural Language Processing (NLP). This paper proposes a novel deep learning based techniques for Sentiment Analysis of Code-Mixed Hindi-English text. The proposed Deep Learning technique uses joint learning from Character and Word features. This paper also implements a technique for preparing corpus for training Word Embedding of Code-Mixed text. This embedding is further used in proposed architecture for robust classification. The proposed approach exceeds baseline accuracy. Additionally this paper shares experiments, which are conducted on the proposed architecture using different loss functions and optimizers.