The purpose of this work is to accumulate information and construct a parts-of-speech tagging strategy for Kannada-language doctor-patient communications. Viterbi decoding and HMM model are utilized to accomplish this. Training data comprises of 20973 Kannada items that have been manually labeled by experts with correct parts of speech tags. Test data comprises of 450 hand gathered and pre-processed observations of doctor-patient discussions from hospitals. In order to determine the best label for the supplied Kannada term, we deployed HMM model with 27 tags. The computation precision was 91.38%. The precision achieved is reasonable compared to existing POS taggers in Kannada. Our distinctive raw data having 450 instances of Kannadalanguage doctor-patient communication was created specifically for this investigation

Hidden Markov Model, Kannada, Machine Learning, Natural language processing, parts of speech tagging