Student: Garfield Jin and Daniel Min
Project: Using Deep Learning to Predict Stock Price Drop | View Poster (PDF)
Institution: Lehigh University
Major: Industrial and Systems Engineering
Advisor: Boris Defourny
Abstract
In this project, we employed deep learning to predict the intraday prices of penny stocks, the common shares of small public companies that trade below $5-per-share in the U.S. market. We focused on penny Stocks because they are too illiquid for Wall Street to scale in, which allows us to avoid competition. In addition, Penny Stocks are very predictable and most of them are called pump and dump, which means when the stock price gets pumped up, it is usually only a matter of time before it loses all its gains. So, we were able to build a realistic model with the sole parameter being the data of prices. The data set we used are the intraday ticks of roughly 400 penny stocks that ran between May.17th and Nov. 29th, 2018. We then cleaned the data and built different models using the Artificial Neural Network to predict the price action of these stocks. Finally, we used our best models to run simulations and trade stocks using Alpaca.
About Garfield Jin
About Daniel Min
Daniel Min is a junior majoring in Industrial & Systems Engineering with a minor in Computer Science. He has previously interned at Quanray Electronics and 5th.AI as a testing engineer assistant and 3-D modeling engineer. This summer, Daniel will intern at TE connectivity at their data analytics department. Daniel’s primary focus within ISE are optimization and machine learning with an interest in Data mining. Daniel enjoys many different kinds of sports such as tennis, soccer and skiing. After graduating, he plans to get a job related to data analytics and artificial intelligence.