Effectiveness of the Philippine Stock Exchange Index (PSEi) as training dataset in forecasting Philippine stock prices using neural networks

• Main Authors: Sumayo, Noriel Kristine Luzanta, Ting, Nico Rafael Ayo
• Format: text
• Language: English
• Published: Animo Repository 2022
• Subjects: Stock price forecasting > Philippines; Stock exchanges > Philippines; Neural networks (Computer science); Mathematics
• Online Access: https://animorepository.dlsu.edu.ph/etdb_math/20; https://animorepository.dlsu.edu.ph/cgi/viewcontent.cgi?article=1019&context=etdb_math
• Institution: De La Salle University

The health of the stock market is considered critical to a country’s economic development. The volatility of stock prices which are influenced by inflation rates, interest rates, tax changes, and other monetary policies, makes the prediction and analysis a very challenging task. With the use of advanced intelligent techniques such as deep learning, we can improve stock market prediction. In this study, we investigate the effectiveness of using the Philippine Stock Exchange index (PSEi) as a training dataset of three artificial neural networks (ANNs), namely, Multilayer Perceptron (MLP), Long-Short Term Memory (LSTM), and Convolutional Neural Network (CNN) in forecasting the daily closing prices of local stocks AbaCore Capital Holdings, Inc. (ABA) and San Miguel Corporation (SMC). Based on the mean squared error (MSE) and mean absolute percentage error (MAPE), the models using MLP with the activation function of hyperbolic tangent (tanh) are the suitable neural network model for both ABA and SMC.