Predicting the stock market has been an unsolvable problem for mathematicians and computer scientists alike for centuries. Thinking he could do better than everyone else, Simon tried to cobble together a machine learning model that could beat everyone else. Using the top news headlines from Reddit, he predicted the opening price of the DJIA with some success (it predicted the correct direction 56% of the time!). This was done using a combination of embeddings, CNNs, and LSTMs.

Log in or sign up for Devpost to join the conversation.