Comparative Study of Rainfall Prediction Modeling Techniques (A Case Study on Srinagar, J&K, India)

Abstract

Prediction of rainfall is one of the most essential and demanding tasks for the weather forecasters since ages. Rainfall prediction plays an important role in the field of farming and industries. Precise rainfall prediction is vital for detecting the heavy rainfall and to provide the information of warnings regarding the natural calamities. Rainfall prediction involves recording the various parameters of weather like wind direction, wind speed, humidity, rainfall, temperature etc. From last few decades, it has been seen that data mining techniques have achieved good performance and accuracy in weather prediction than traditional statistical methods. This research work aims to compare the performance of few data mining algorithms for predicting rainfall using historical weather data of Srinagar, India, which is collected from http://www.wundergrounds.com website. From the collected weather data which comprises of 9 attributes, only 5 attributes which are most relevant to rainfall prediction are considered. Data mining process model is followed to obtain accurate and correct prediction results. In this paper, various data mining algorithms were explored which include decision tree based J48, Random forest, Naive Bayes, Bayes Net, Logistic Regression, IBk, PART and bagging. The experimental results show that J48 algorithm has good level of accuracy than other algorithms.