Comparative Analysis of Machine Learning Algorithms for Rainfall Prediction

Abstract

AbstractIndia is an agricultural nation and its economy is based primarily on agricultural productivity and precipitation. Prediction over rain is expected and appropriate for all farmers in order to evaluate crop productivity. Rainfall forecast is the application of science and technology to estimate atmospheric status. The precipitation for efficient use, crop production and preplanning of water systems should be accurately calculated. The prediction of rains can be done with the data mining classification task. The performance of various techniques depends on the representation of rainfall data, which includes long-term (month) pattern as well as short-term (daily) pattern representation. It is a challenging task to pick an effective strategy for a specific period of rain. This article focuses on few prevalent rainfall prediction data mining algorithms. Naive Bayes, K-Nearest Neighbor Algorithm, Decision Tree, are some of the algorithms compared in this article. The Delhi NCR region weather data collection from 1997 to 2016 was collected. The approach can be evaluated for better rainfall forecast accuracy. Experimental results demonstrate Decision Tree Classifier (DTC) is powerful in the extraction of rainfall prediction. Applications of Decision Tree Classifier will provide accurate and timely rainfall prediction to support local heavy rain emergency response.KeywordsNaive BayesK-nearest neighbor algorithmDecision treeRainfall predictionData miningMachine learning