ANALYSIS OF NASDAQ COMPOSITE INDEX PRICE PREDICTION USING WAVELET NEURAL NETWORK METHOD

• Main Author: Ghifari, Faisal
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/70320
• Institution: Institut Teknologi Bandung

Econophysics is a discipline that applies models and concepts derived from physics to economic and financial phenomena. One of things that studied in econophysics is stock price prediction. Stocks are securities that signify the ownership of a fraction of a company and are used as a form of investment. Stock prices continue to change as they adjust to the market. Investors can gain profit by selling stocks at higher prices than its purchase price. This makes stock price analysis and prediction important to gain profit and avoid losses when investing in stocks. A method that can be used to predict stock price is wavelet neural network. Wavelet neural network is a machine learning method that combines wavelet transformation and artificial neural network. This thesis aims to analyze performance of the wavelet neural network method in predicting the close price of four stocks that are part of the Nasdaq Composite stock index. Wavelet neural network used is a combination of Discrete Wavelet Transform (DWT) and Long-Short Term Memory (LSTM). As a comparison of wavelet neural network, LSTM method is used. Close price of the four stocks that are used are from April 2020 until March 2022. This data divided to train data and test data. The result shows that wavelet neural network outperform LSTM.