Comparative Analysis of LSTM Models for Forecasting Millet Production in Maharashtra

Intensive emphasis has been put on the growth and production of millets during the last couple of years. It's their high intrinsic qualities and lower requirements that the government is trying to promote the consumption and production of millets through various awareness missions. Due to the limitations in data availability, it becomes challenging to fine-tune the models in the context of the available data. In this paper, we have performed a comparative analysis between two different variants of Long Short-Term Memory models for forecasting the production of millets in the Indian state of Maharashtra. Based on the parameters generating optimal forecasts, we have forecasted the production of Ragi, Bajra, Jowar, and other minor millets. An analysis of both granular and aggregate level forecasts has been carried out to study and explore the trends in the data and forecast production for the future.