A flower pollination algorithm based Chebyshev polynomial neural network for net asset value prediction

Deriving relevant features from historic financial data and forecasting it accurately and impartially is an emerging and predominant field of research in the vast domain of finance. Mutual fund is a structured investment instrument in economic market and the measuring instrument required to price it is called as Net Asset Value (NAV). For NAV prediction, Artificial Neural Network has been immensely utilized in the past due to its adaptive learning, robustness and great ability to identify and handle hidden nonlinear patterns. In this study, a Chebyshev Polynomial Neural Network (CPNN) has been proposed for one day ahead prediction of three different NAV data set belonging to three leading Indian financial houses. The controlling parameters of the network are estimated by a computationally intelligent meta-heuristic algorithm known as Flower Pollination Algorithm (FPA). It is a nature inspired algorithm, motivated by the pollination process of flowering plants with very few control parameters and rapid convergence rate. The efficacy of the proposed forecasting model is also examined by a comparative analysis of training of CPNN with other optimizing algorithms like Differential Evolution and Particle Swarm Optimization (PSO) on the same set of data collected over the same period of time. The convergence plots obtained during the training of the models and the different error metrics values calculated during their testing phase of the study showcase that the proposed CPNN-FPA model clearly outperforms the other two experimented predictive models.