Investigating Effective Factors on Sugarcane Production Performance to Increase the Production of Sugarcane Using Data Mining

Sugarcane is an important industrial -agricultural crop in country. Considering the high cultivation level of this product in Khuzestan province and the high volume of data stored in these mechanized agricultural units, it is necessary to have a tool to process stored data. The data mining technique is well equipped to provide sugarcane manufacturers with the necessary information and patterns in modeling the yield of sugarcane. One of the most practical of these algorithms is decision trees. The main objective of this research is to predict yield sugarcane and to evaluate the factors affecting it using decision trees CART and CHAID. The present work was an analytical study conducted on a database containing 13211 records. Data were obtained from farms of Amir Kabir Agro-Industry, during 2013-2016. Data analysis was performed using IBM modeler software version 14.2 by CRISP methodology. The accuracy of decision tree (CART and CHAID) on the training data and test data were 90, 81, 85 and 79 percent, respectively. The results of this research can be used for sugarcane production and cultivation industries in Khuzestan Province in order to evaluate and optimize the sugarcane production process and predict the yield of sugarcane.