Simulation Based Avalanche Prediction Model using Adaptive Neuro Fuzzy Inference System

Abstract

The main aim of this study is to develop an avalanche prediction method, based on the adaptive neural-based fuzzy inference system (ANFIS). An ANFIS methodology is applied to the sample weather data inputted in the Matlab through Microsoft’s excel sheets. Application is given for atmospherically time series modeling, and then the sample data is used for training data sets of the ANFIS. It is seen that the extension of input and output data sets in the training stage improves the accuracy of forecasting by using ANFIS.

The time required for the analysis and prediction of an extremely volatile event like avalanche needs to be reduced to the minimum. This is particularly critical because of the extremely fast and highly uncertain nature of the event itself. Another peculiar nature of such predictions is that these have to be based almost entirely on the long and intermediate-term data/information available. Both the above-mentioned factors favour adoption of such techniques of automated analysis, which are fast, accurate, and employable even under uncertain voids of information.

Apart from empirical and statistical methods, one of the highly promising techniques for developing a practical model for prediction of avalanche is that based on rules. The process of defining a highly uncertain phenomenon like the avalanche at such high resolution, and thereafter, framing extensive rules for all the possibilities is likely to make the system extremely complex, and therefore, unmanageable in many ways. The present study attempts to simplify this problem by proposing a simpler and better technique using an algorithm based on fuzzy logic.