TECHCOMB

Voume 3 : Issue 1

Genetic Algorithms Based Fuzzy Time Series Prediction for Water Table Elevation Fluctuation

Authors : Shilpa Jain.,Dinesh Bisht.

image

Abstract:

Fuzzy time series is a powerful forecasting technique. It performs well when no statistical trend or cycle is available in the data. Fuzzy time series has been applied to the prediction of enrollment of students, temperature, stock indices and many areas of environment and ecology. The area of prediction of water table elevation fluctuation is always an area of interest to researchers due to very high consumption of water. Genetic Algorithms is an optimization and search technique. In this research article a combined fuzzy and Genetic Algorithm approach is used for prediction of water table elevation fluctuation. In proposed method binary GA is used to adjust the length of fuzzy intervals and rule base is constructed with the help of a method proposed by Wang and Mendel. This method is implemented on three models of prediction of water table elevation fluctuation. Thousand generations were explored to state that Coefficient of Determination (R2) for Fuzzy model varies from 60.8% to 90.8% whereas Fuzzy combined with GA improves it from 89.9% to 91.0%.