Algo-stocks: The new figure in stock price prediction and strategic trading
Date of Publication
Bachelor of Science in Management of Financial Institutions
Finance and Financial Management
Ramon V. Del Rosario College of Business
Financial Management Department
Mar Andriel Umali
Defense Panel Chair
Defense Panel Member
Investing in the stock market has been around since the 1500s, therefore, there are millions of stock traders all around the globe. Through the decades, fundamental and technical analysis were the key methods used. However, with the rise of technology and wide access to information, the effectivity of these methods may have changed.
This study aimed to find out whether machine learning algorithm, k-nearest neighbor (k-NN), is more accurate model than technical analysis, moving average (MA), in predicting next day closing stock prices. Root mean square error (RMSE), mean percentage error (MPE), and average difference (AD) were used as back testing models, and the researchers pushed the envelope even further and conducted a trading simulation using the next day forecasted prices. Based on the results, it was found out that k-NN was the better forecasting model than MA in terms of RMSE and AD. However, MA was the more profitable model when used in the daily trading strategy.
Overall, this study aimed to explore the realm of machine learning algorithm being applied in the stock market, and aimed to show an option to traders, who are currently using MA in their trading strategies, to use k-NN in conjunction with other indicators to make better price predictions and generate more profits in the stock market.
Archives, The Learning Commons, 12F, Henry Sy Sr. Hall
ix, 86,  leaves : illustrations ; 28 cm.
Stock price forecasting--Philippines
Cruz, M., Ong, A. T., Ong, J. S., & Ong, S. G. (2017). Algo-stocks: The new figure in stock price prediction and strategic trading. Retrieved from https://animorepository.dlsu.edu.ph/etd_bachelors/9033