Author(s): K. G. Arunya; M. Krishnaveni

Linked Author(s):

Keywords: Machine learning; Precipitation; Crop yield; Participatory rural appraisal

Abstract: Machine learning (ML) is an important tool in decision-making and many ML algorithms are used in crop yield prediction. ML algorithm predictions are based on various dependent factors. In this study, climate variabilities such as precipitation and crop yield are considered as the dependent factors. This paper shows a micro-level study which concentrates on the Paravanar River Basin which is situated on Cuddalore District, Tamil Nadu, India. There are more than 40 crops grown in this region and the main source of irrigation is the north-east monsoon rainfall. Since the hypothesis is tested between the annual rainfall and the crop yield, the crops which are cultivated throughout the year are chosen for analysis. They are cashew nut, coriander, sugarcane, sweet potato and turmeric. The PERSIANN-CDR data from National Centres for Environmental Information—National Oceanic and Atmospheric Administration was used as the precipitation data and the agriculture crop production data was downloaded from Ministry of Agriculture and Farmers Welfare of India. The datasets were tested using the following ML algorithms: logistic regression, decision tree classifier, random forest classifier and XGBoost classifier. Implementation of ML-derived results can only be done through participatory approaches. Therefore, participatory rural appraisal was done with the farmers and the villagers to assess the willingness of implementation and it shows positive results. Hence the main aim of the study tests the hypotheses on dependency of crop yield on precipitation on a regional scale.

DOI: https://doi.org/10.1007/978-981-97-6009-1_33

Year: 2022