Are the Adivasi Rice Growers of Sub-Himalayan North Bengal Efficient Enough? Role of Machine Learning in Farm Business After COVID-19

Abstract

The COVID-19 pandemic and the consequent lockdown have disrupted the farming communities across the planet; however, the indigenous communities such as the Adivasi people of tea garden areas in India have been hit hard. It becomes more crucial when we talk about marginalized communities. In India, the Adivasi tribe found in the sub-Himalayan North Bengal cultivates folk rice that needs immediate energy-efficient measures as the production has been increased after receiving the geographical indication tag. Energy efficiency estimation often applied a two-step data envelopment analysis model in agricultural production. However, in most of the previous articles, the applications discussed the factors affecting energy use efficiency with rare studies on efficiency prediction. In this article, first, data envelopment analysis was used to estimate the energy efficiency of rice growers, and in the second stage, extreme gradient boosting, a state-of-the-art machine learning algorithm, was employed to derive the key leading efficiency determinants. The findings revealed wide variation among efficient and inefficient rice growers in the first stage and derived the most salient factors predicting energy efficiency in the second stage. The optimal use of energy inputs combined with effective education, better credit delivery mechanism, arable land availability and years of farming experience provided improvement for the future energy use efficiency of the Adivasi farmers. Further, the novel application of extreme gradient boosting as a machine learning algorithm in energy efficiency evaluation suggests decision-making solutions with a prediction accuracy of 80.91%. Moreover, this study aims to assist future researchers in examining and predicting the key leading determinants to affect energy utilization.