Predicting the Quantity of Agricultural Wastes Using WEKA ‎Application

Abdulwahed M. Aboukarima*(1,2) Samy G. Hemeda(2) Mohamed S. El Marazky(2) Riham El_Oliemy(2) and Ibrahim S. Tabash(1)

(1). Department of Agricultural Engineering, Faculty of Food Sciences, King Saud University, Riyadh, Saudi Arabia.

(2). Agricultural Engineering Research Institute, Agricultural Research Centre, Giza, Egypt

(*Corresponding author: Dr. Abdulwahed M. Aboukarima. E-Mail: aboukarima@gmail.com)

Received: 29/03/2020                               Accepted: 06/06/2020

Abstract

In this study, a comparison was made between data mining tools in WEKA open source computer application to predict the amount of agricultural wastes in Egypt from the year 2006 to 2016. The amount of wastes was calculated according to the method presented in the Arab Organization for Agricultural Development (AOAD). The results showed that the lowest amount of wastes was 13.57 million tons/year and the largest amount was 17.90 million tons/year. The results revealed that the data mining tool called Decision Table was the best tool for predicting the amount of agricultural wastes. The predicted lowest and the largest means were 2502.17 and 3579.66 thousand tons/year, respectively, while the mean actual value in the test data set was 3300.06 thousand tons/year. The study recommended that adequate attention was directed to forecast agricultural wastes amounts, as these wastes can be managed and utilized in the production of energy, compost or animal feed.

Key words: Machine learning, Agricultural wastes, Modeling, Environment.

Full paper in Arabic: PDF