Rainfall prediction is indeed in growing unpredictable natural phenomenon. Using the history of data associated with rainfall in a selected scope to predict the upcoming scenario is unit of the widely used methods in weather forecasting mainly in rainfall prediction. Processing the data sets collected from the past to forecast weather involves a popular data mining methodology known as flocking. Here we consider a latest approach of segmentation methodology known as Reinforced K-Means Hadoop MapReduce methodology (EKM-HMR) where a modified Kmeans map-reduce computation aids to reduce the distance between the flocks. This reinforced methodology works properly for all data sets as the calculation metric such we have considered here would be generally applicable for all kinds of datasets. This makes the computation go with applications such contain varying data set types as inlet which aids the flocks to process the datasets, even more, faster than usual which reduces the execution time and in turn raises precision. The outcome evaluation of the contemplated methodology is highlighted by comparing the precision with K-means flocking computation. Results from the evaluation criteria reveal such the EKM-HMR performs better in predicting the weather mainly rainfall.
Volume 12 | 03-Special Issue