Rice yield prediction using a support vector regression method

• Main Authors: Ratchaphum Jaikla, Sansanee Auephanwiriyakul, Attachai Jintrawet
• Format: Conference Proceeding
• Published: 2018
• Subjects: Computer Science; Engineering
• Online Access: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=52949132203&origin=inward; http://cmuir.cmu.ac.th/jspui/handle/6653943832/60290
• Institution: Chiang Mai University

Rice yield prediction is the procedure to predict the rice grain weight. The objectives of the procedure are finding out whether the location is appropriate to grow rice, and reducing any risk in the investment of rice yield production. There were many researchers trying to find the precise results of rice yield prediction, however, the proposed methods are complicated and unique. This paper, therefore, is aimed to develop rice yield prediction procedure using the Support Vector Regression method (SVR), one of the most widely used techniques in data prediction. The prediction method in this paper is divided into 3 phases, i.e., soil nitrogen prediction, rice stem weight prediction and rice grain weight prediction. We compare the results with the commercial software, i.e., DSSAT4 program implementing Crop Simulation Model (CSM-Rice simulation model). The results indicate that our method is comparable with that of the CSM-Rice simulation model. The error from our model is also in the acceptable range. ©2008 IEEE.