Language Identification and Translation of English and Gujarati code-mixed data

Code-mixing is a growing field of research in the domain of Natural Language Processing. Communication on social media involves code-mixed texts, colloquial language, and variations in spellings. An amalgamation of different languages in the conversation results in transliteration and romanization. Presently there is a deficit amount of open-source resources for code-mixed English-Gujarati data, while the amount of such data is exponentially increasing every-day with growing social-media users. In this paper, we present our creation of a linguistic resource for English and Gujarati. We also present our approach of language identification and normalization of data along with the translation of the transliterated text into native language form.