Research on Efficient Landslide Prediction Approaches using Machine Learning Techniques

Abstract

A landslide is a condition in which a huge amount of rock particles slide or break off down a slope, resulting in great natural and physical loss in addition to the lives of many people. In large parts of the world, massive damage is caused by landslides. The utility of remotely sensed images is used for landslide detection, mapping, prediction, and assessment round the world. This systematic analysis might also make contributions to better expertise the considerable use of remotely sensed records and spatial evaluation techniques to conduct landslide research at more than a few scales. The machine learning algorithms in particular ANN and SVM are used as soft computing techniques for landslide prediction. The accuracy obtained from SVM is 91.78% and with ANN 93.38%. In India landslide is famous phenomena of Himalayan location, Western Ghats and southern Nilgiris Mountains. Such losses must be avoided if right perception tool is available that would notify about the event in boost. With the use of proposed soft computing techniques this paper projects unique landslide prediction techniques with cognizance on western India.