Integrating Fundamental and Technical Analysis of Stock Market through Multi-layer Perceptron

We use Multi-layer Perceptron and propose a hybrid model of fundamental and technical analysis by utilizing stock prices (from 2012-06 to 2017-12) and financial ratios of Technology companies listed on Nasdaq. Our model uses data discretization and feature selection preprocesses. The best results are obtained through topology optimizations using a three-hidden layer MLP. We examine the predictability of our hybrid model through a training/test split and cross-validation. It is found that the hybrid model successfully predicts the future stock movements. Our model results in the greatest average directional accuracy (65.87%) compared to the results obtained from the fundamental and technical analysis in isolation. The numerical results provide enough evidence to conclude that the market is not perfectly efficient.