Stock Market Forecast

Web App Example for Google Stock

My Project

The goal of my project was to forecast the shot-term growth of a stock using machine learning. I made a WebApp in which you can enter the ticker of any stock and get its forecasted growth.

How I built it

My webapp automically retrieves the latest data about the stock using an API. It then calculates technical indicators about the stock such as momentum and feeds it to my model. I experimented with many variables and kept the best predictive variables. I chose an LSTM model to try to detect patterns/correlations within the data and did some time series analysis.

Accomplishments & Continuation

My model is able to predict prices quite accurately the prices with a mean squared error of 100. It did, however, not perform consistently well to forecast the movement of a stock price. Continuing on, I'd definitely prefer to train my model more with a focus on the movement on the market along with more data.

Learning Experience

Throughout this project, I learned the inner workings of the LSTM neural network. I've learned about a lot of statistical concepts such as time series analysis. Finally, I've gained solid knowledge in the world of finance.