Automatic Normalization of Word Variations in Code-Mixed Social Media Text

Social media platforms such as Twitter and Facebook are becoming popular in multilingual societies. This trend induces portmanteau of South Asian languages with English. The blend multiple code-mixed data has recently become research communities for various NLP tasks. Code-mixed consist anomalies grammatical errors spelling variations. In this paper, we leverage the contextual property words where different variation share similar context a large noisy social text. We capture variations belonging to same an unsupervised manner using distributed representations words. Our experiments reveal that preprocessing dataset based on our approach improves performance state-of-the-art part-of-speech tagging (POS-tagging) sentiment analysis