Analysing Word Stress and its effects on Assamese and Mizo using Machine Learning

Stress identification is an important problem in speech processing, which aims to convey special attention to the listeners. Stressed words or syllables result in changing the meaning of a sentence. In the case of tonal languages, stress identification is essential to understand how stress over words may affect the tone. This work is an attempt to identify the effects of stress on tones in Mizo and intonation in Assamese. It also aims to analyze the co-articulatory effects of stress and tones in Mizo and stress and intonation in Assamese with the help of an audio dataset. Although a few works are available to identify the effects of tones and stress, for Indian languages especially, in North East Indian languages which are extremely low in resources, to the best of our knowledge, no such work is available. For our work, we have considered one tonal language, i.e., Mizo or Lushai, spoken in and around the state of Mizoram, and one non-tonal language, i.e., Assamese, spoken in and around the state of Assam in India.