Stock Market Prediction in Malaysia using Convolutional Neural Network

• Main Author: Nazurah Batrisyia Syaurah, Binti Mohamad Najib
• Format: E-LPTA
• Language: English; English
• Published: Universiti Malaysia Sarawak (UNIMAS) 2019
• Subjects: BF Psychology; H Social Sciences (General)
• Online Access: http://ir.unimas.my/id/eprint/27548/1/Stock%20market%20prediction%20in%20Malaysia%20using%20convolutional%20neural%20network%20%2824%20pgs%29.pdf; http://ir.unimas.my/id/eprint/27548/2/Stock%20market%20prediction%20in%20Malaysia%20using%20convolutional%20neural%20network%20%28fulltext%29.pdf; http://ir.unimas.my/id/eprint/27548/

Stock market is an important element in business as it may contribute to profit or loss in a company. Therefore, a good prediction of the direction of the stock market is crucial in order to avoid a big loss in a company's business. The aim of this study is to obtain higher accuracy of stock market prediction in Malaysia using deep learning, specifically Convolutional Neural Network (CNN). CNN is commonly used for image recognition. Besides this study also determined to find out the capability of CNN to predict 1 D-CNN dataset. The dataset used is from the year of 2006 to 2018 and obtained from Yahoo! Finance. The steps involved were data processing, data training and testing and evaluation. For data processing, normalisation technique is applied to avoid large gap between the dataset. Training and testing used I D-CNN algorithms and the evaluation analysis is by calculating the accuracy, precision, recall and F 1-score. The findings were then compared with multi-layered perceptron (MLP) and the result showed CNN scored higher to be compared to MLP by 83% and 76% respectively.