DETECTING PRICE MANIPULATION IN UMA POLICY USING NEURAL NETWORK MODEL CASE OF INDONESIAN CAPITAL MARKET

• Main Author: Adhinugroho, Triasto
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/40967
• Institution: Institut Teknologi Bandung

as harmful practices in global share markets. We investigated the characteristics of the stock price manipulation. Two manipulation techniques were studied: pump and dump, spoof trading. Due to the new technology, machine learning algorithms now supporting to detecting stock price manipulation. In this study we use Indonesia Stock Exchange intra-day data and we investigated stocks involved in the Unusual Market Activity (UMA) Announcements during 2008-2017, is there any connection between the UMA and the stock price manipulation. We use Neural network model to detect stock price manipulation for each manipulation techniques, after that we want to detect price manipulation, so we took randomly 30 stocks and input it into 2008- 2017 model and from the output we got the UMA policy can’t detect price manipulation. From this study we discover that UMA can be proxies to detect stock price manipulation, but for long term the announcement cant be a proxy anymore because of technology and science are continuously evolving , the Indonesia Stock Exchange should look for better consideration and evaluation in assessing and publishing UMA announcements