Use of Machine Learning Tools in Geographic Information Systems Resource Planning Applications

This paper is the result of a search for a method for machine learning in Geographic Information Systems applications. GIS is basically a complex database system that stores and manipulates graphics map data with analytical capabilities.Machine learning can be employed in extracting information from the graphical and attribute data sets already stored in a GIS by means of various machine learning techniques, GIS is used to generate spatial variables data as an input to a machine learning tool. For the application of Neural Networks (NN) method, Weka data mining suite has been used with multi-layer perceptron algorithm. The proposed solution requires 13 attributes including direct weather inputs (temperature, relative humidity, wind speed and rain) which are used as input within a geographic grid (area) system, for predicting forest fires. It creates a model of prediction on the data obtained from the past occurences of forest fires. The results which have geographic references with a grid system can be used in preparing "Risk Mapping", damage assessment and planning of the resource allocation in fighting forest fires.