A Prognostic Approach For Precipitation Forecast Using Naive Bayes Algorithm

Prognosis of precipitation is essential and challenging job for the researchers. Predictingaccurate precipitation is vital for water resource management in agriculture field and disastermanagement and its allied sectors. Rainfall prognosis is a type of weather forecasting whichinvolves taking down the various aspects of conditions. In the recent past, it has been shown thatdata mining techniques along with machine learning algorithms have better performance andprediction than conventional statistical methods. In this work, we have used Naive Bayes techniquefor the rainfall prediction. The weather information of Srinagar, India, is collected fromhttp:///www.wundergrounds.com website. Five (5) most significant weather attributes forprecipitation prognosis are selected from nine (9) attributes. Analysis and testing the prediction byNaive Bayes Algorithm is performed by comparing the rainfall prediction results of actual data itemfor a particular day using Naive Bayes Algorithm with the observed data item of the collectedweather data set. The results of the Naive Bayes Algorithm are found to be quite accurate andacceptable