Abstract:
Agriculture is the main economic activity in Rwanda and tea is major cash crop in
Rwanda. There has been extensive research on prediction of tea production but most of
the methods applied were the traditional statistical analyzes with limited prediction
capability. Data mining algorithm models, linear regression, K-Nearest Neighbor (KNN),
Random Forest Regression, Extremely Randomised Trees are discussed in this study to
identify critical features in different domains to facilitate accurate prediction of tea
production in Rwanda. In this study also, I identified different factors which are strongly
associated with tea production and developed data mining models for predicting tea
production using training and test data from National Agricultural Export Development
Board (NAEB) 2010-2019. The findings reveal that random forest is the best model
among the others to predict tea production in Rwanda.
