TRADING DECISION PREDICTION OF BANK CENTRAL ASIA STOCK WITH TECHNICAL ANALYSIS USING NEURAL NETWORK MODEL

• Main Author: Jordan Enrico, Misael
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/47820
• Institution: Institut Teknologi Bandung

The practice of stock trading is one of the way one can invest their money into. Stock trading is done only if its price is agreed upon by both the buyer and seller. Generally, stock price can be characterized as having a mild fluctuation caused by imbalance between supply and demand. Having that said, in-depth analysis regarding stock price movement is required so that one can maximize profit from stock trading. Two analyses that are widely known by stock traders are technical analysis and fundamental analysis. However, techincal analysis is a better short-term analysis due to its responsiveness in capturing short-term price movement than its counterpart. Technical analysis relies on mathematical calculations known as technical indicators. Previous studies have made an attempt to model stock price movement by predicting latter stock price as well as trading decision: a decision whether on a certain day one should buy, sell, or hold stocks. Model with its output in form of trading decision simplifies the problem of price movement prediction. Many types of model has been utilized to obtain good prediction on stock market data with one of them being neural network model. Neural network architecture imitates biological neural system in organism. This allows information to be passed through several neurons, resulting in complicated calculation. This model is able to capture complex relationship between its input variables compared to many other model. In this thesis, trading decision well be determined using one of several neural network architectures known as ModAugNet with technical analysis indicator being taken into account so that it can be applied to predict trading decision of PT Bank Central Asia Tbk stock.