A low complexity evolutionary computationally efficient recurrent Functional link Neural Network for time series forecasting

The paper presents a low complexity recurrent Functional link Artificial Neural Network for predicting the time series data like the stock market indices over a time frame varying from one day ahead to one month ahead. Further an adaptive bioinspired Firefly algorithm is adopted here to find the optimal weights for the recurrent computationally efficient functional link neural network (RCEFLANN) using a combination of Linear and hyperbolic tangent basis functions. The performance of the Recurrent computationally efficient FLANN model is applied for the prediction stock prices of Standard & Poor´s 500 (S&P500) and NIKKEI 225 data sets providing significant accuracy. Also another time series data like the electricity price of Pennsylvania-New Jersey-Maryland (PJM) energy market has been considered for weekly electricity price prediction using the proposed approach with satisfactory results.