Association Rule Mining for a Climatic Condition Based Recommender System: A Cinnamon Cultivation Case Study from Galle, Sri Lanka

Abstract

Weather prediction has become complex due to ever-changing weather conditions. Effective analysis of the weather requires the identification of influential factors in changing weather conditions. Several machine learning techniques have been applied for whether predictions during the past few years. Linear regression, functional regression, neural networks, classification and regression trees, k-nearest neighbours, and Naïve Bayes are frequent among them. In this work, we thoroughly analyze machine learning techniques, especially the usage of recommender systems for weather prediction. Further, analyze meteorological data of Galle district in Sri Lanka using suitable techniques to find hidden patterns by transforming the historical information into usable knowledge. The main objective of this paper is to propose a recommender system for Cinnamon cultivation in the Galle district to predict future weather conditions based on association rule mining. This can further be used to identify the best sowing dates in the future for not only cinnamon, but also tea and other cultivations in the district.