RESEMBLANCE OF RAIN FALL IN BANGLADESH WITH CORRELATION DIMENSION AND NEURAL NETWORK LEARNING

Rain fall and Temperature are undoubtedly two important factors that balance water in the environment. Adequate study of the rain behavior helps to forecast it. The time series obtained from different stations of the country throughout the several years are collected and analyzed. The dynamics of rain fall time series is analyzed with Correlation Dimension (CD), to characterize the several zones of Bangladesh. In addition, a Neural Network (NN) predictor model was designed to realize complexity of rain fall. The authors have found the interesting similarity between CD and NN predictor. The findings are useful in explaining why several zones show behavioral regularity and change.