Building an Information Retrieval Apparatus for Kannada Language: A Rule-Based Machine Interpretation System with Noise Suppression

Abstract

India is a home to various vernaculars. Due to the widespread usage of the Internet, the amount of information available is huge, thus people tend to bend toward these available resources for the information retrieval for various vernaculars. Indian dialects give charming four particular vernacular families, to be specific; those Indo-Aryan, those Austro-Asiatic, those Tibeto-Burman, and the Dravidian. Kannada belongs to the Dravidian family. The main focus of this paper is interpretation of Kannada language in particular for the people familiar in English to get access to the required information without much difficulty. Suitably, it is fundamental to build an information retrieval (IR) instrument or device for the Kannada composite. Keeping in mind the end goal to build up an IR apparatus encompassing the Kannada dialect, the specifics of the dialect are to be recognized. This paper gives a rule-based approach which additionally strips fourteen diverse significant portrayals of postfixes (pratyaya in Kannada), along with the subclasses by interpreting morphophonemic assurance and involving a zero-shot noise suppressing algorithm for the texts or document(s) encompassing images along with them.