Abstract
The correct and timely forecast of the Indian summer monsoon Intraseasonal Oscillations (ISOs) is very important. It has great impact on the agriculture and economy of the Indian subcontinent region. The applicability of Genetic Algorithm (GA) is demonstrated for nonlinear curve fitting of the inherently chaotic and noisy Lorenz time series and the ISO data. A robust method is developed for the very long‐range prediction of the ISO using a feed‐forward time delay backpropagation Artificial Neural Network (ANN). Using an iterative one‐step‐ahead prediction strategy, five years (120 pentads) of advanced prediction is made for the ISO data with good forecast skill. It is shown that a hybrid GA‐ANN model may be used as an early forecast model followed by ANN only model as a more reliable model.

This publication has 31 references indexed in Scilit: