Title: Predicting the soil profile through modified regression by discretisation algorithm for the crop yield in Trichy district, India
Authors: M.C.S. Geetha; I. Elizabeth Shanthi
Addresses: Department of Computer Applications, Kumaraguru College of Technology, Coimbatore, Tamil Nadu, India ' Department of Computer Science, Avinashilingam Institution for Home Science and Higher Education for Women, Coimbatore, Tamil Nadu, India
Abstract: Agriculture seems to be the most prime segment of the Tamil nadu State financial system, as around 70% of the residents are occupied in agriculture and associated activities for their source of revenue. But still, the state faces problems of resource-poor farmers, fragmentation of holdings, reliance on monsoon rain, and poor soil productivity. We have taken up the issue of predicting the yield level by analysing the soil characteristics for the particular soil. As gathering the information from an enormous volume of data is practically a complicated task in the recent circumstances, we have concentrated only on Trichy district and the soil data samples have been collected. The modified regression by discretisation was used for training and testing of the data sets and the results were explored in order to obtain the crop yield prediction for the soil. This study proves it's best for training soil data for classes' prediction.
Keywords: agriculture; classifiers; regression by discretisation; soil; Trichy district; India.
International Journal of Grid and Utility Computing, 2018 Vol.9 No.3, pp.235 - 242
Received: 05 Dec 2017
Accepted: 05 Feb 2018
Published online: 10 Aug 2018 *