Abstract:

Even though a number of stock market forecasting studies are done related with hybrid Artificial Neural Network (ANN) models, no standard procedures are available in the literature for each stock. This causes a growing interest in using metaheuristic for the designing of appropriate ANN architecture. Therefore, this study used ten different metaheuristics including Ant Lion Optimization (ALO), Bird Swarm Optimization (BSA), Differential Evolution (DE), Grey Wolf Optimization (GWO), Moth-Flame Optimization (MFO), Multi-verse Optimizer (MVO), Particle Swarm Optimization (PSO), Simulated Annealing (SA), Weighted Superposition Attraction (WSA), and Firefly Algorithm (FFLY) to improve the performance of the ANN models. Proposed hybrid ANN models lead to significant opportunities to forecast stock market more effectively. Based upon results of performance measures, we also expect hybrid ANN models provide a remarkable solution for other forecasting problems.

Keywords: