Rainfall Events Predication Using Data Clustering an Application to North-Khorasan (Iran)

This paper Precipitation rate forecast is one of the most important factors in managing the water resources of a country. In many countries, precipitation rate is directly related to economic growth and even the industries are highly dependent on it. Modeling and prediction of natural phenomena are very difficult tasks, owing to high complexity and uncertainty. Numerous highly dynamic factors contribute to predict these processes. In this paper, efforts have been made to forecast precipitation in Northern Khorasan province (in Iran) using different models of data clustering. The required data are received from the synoptic station of the Bojnourd town. The data correspond to the past 32 years and they enjoy great accuracy and precision. These data comprise temperature (minimum, maximum, average), relative humidity, wind velocity, and daily precipitation rate. K-means, fuzzy c-means, and hierarchical methods were used to cluster data. The results of the three methods were compared and the fuzzy c-means method could forecast the precipitation rate in Northern Khorasan province with an accuracy of 96.52%.