This presentation explores the use of Convolutional Neural Networks (CNNs) for stock market prediction, focusing on how image conversion of OHLC data, model complexity, and label types impact forecasting accuracy. The study highlights the benefits of input normalization, deeper CNN architectures, and the superior performance of categorized labels over continuous return rates, offering insights into improving AI-driven stock forecasting.
Kihyung Kim, PhD is an Assistant Teaching Professor in the Management Department at the Trulaske College of Business, University of Missouri. He holds a PhD in Industrial Engineering from Purdue University and has extensive expertise in artificial intelligence, deep learning, business analytics, and supply chain management. Dr. Kim has published in journals like Clothing and Textiles Research Journal and Resources Policy, presented at conferences such as INFORMS Annual Meetings, and integrates AI tools like ChatGPT into business education to enhance analytics and decision-making skills.
