Optimized scenario of temperature forecasting using SOA and soft computing techniques

Abstract

Weather forecasting at a given instant of time and location is a challenging activity as its data are continuous, highly intensive, multidimensional, and dynamic in nature. This paper presents an approach for maximum temperature forecasting over a given period of time using the service-oriented architecture (SOA) and soft computing techniques. SOA is used for collecting data of a particular location using the principles of SOA, i.e., reusability, interoperability, and composability. Large number of attributes of weather dataset gathered with the help of SOA concept can be curtailed sing one of the soft computing techniques, i.e., rough set theory (RST). The RST technique works by finding the relevant attributes and eliminating the irrelevant attribute which are not essential. The residual attributes are used to forecast temperature based on artificial neural network (ANN) technique. RST technique has been applied to improve the performance of ANN computationally as well as by its accuracy.