Abstract

Abstract: India is an agriculture based country. Farming andfarm related business are mainly focused for earning by major population.   Farmer has to make decisions manually everyyear for various factors such as crop to be cultivated, choosing a fertilizerquantity as per soil properties. The decisions made are not accurate as it doesnot consider soil properties. To overcome this problem, researchers haveanalysed the agricultural data for various parameters. In this dissertationwork, we are going to use agricultural data which will be region specific. PAM,CLARA, modified DBSCAN data mining culturing techniques are used in literature.We are going to use machine learning techniques namely multiple linearregression, SVM, KNN and ANN for crop yield prediction. Predicting thoseresults on parameters of soil property like NPK values, temperature andrainfall. From this analysis, farmer can get accuracy and clarity to getdecision for crop to be cultivated to get better product yield whileconsidering environmental factors, weather conditions, soil types and season.

Keywords-Machine learning, Data mining, Agriculture