New Machine Learning based Approach for Predictive Modeling on Spatial Data

Image classification domain has been an area which has attracted a lot of researchers over past years. Many classification methodologies for spatial image datasets has been developed. Artificial intelligence based approaches are getting popular now a days for getting the image classification task done in more efficient and correct way. The prime goal is to develop a classification mechanism which can handle the uncertainity in more efficient manner. This paper presents a novel approach for classification based predictive modeling technique for spatial data sets. Advanced machine learning concepts like RST are employed in our algorithmic procedure, which gives a major advantage of reduction in dimensionality of data in efficient way. Experiments are performed on Bhuvan NRSC geospatial repository datasets and simulation results are given.