Our inspiration for this product was based on our fascination for automated financial practices. With the rise of machine learning being ubiquitous throughout the financial sector, we wanted to develop our own model to predict the value of stocks and help day traders assess the risk coming with investment.

What it does

Our web application accepts user inputs to evaluate options and stocks. For example, you can choose any stock trading on NASDAQ and select how many days you want to sell it in and you will receive a predicted value of the stock. Moreover, for options, you can include your buying/strike price as well as when you want to execute the option, and you will receive a valuation on that as well.

How we built it

For this product, we implemented a linear regression model using python machine learning frameworks such as SciKit learn to predict future values of the stock. We trained the data using the yahoo finance python library and received a ninety-eight percent accuracy with this model. Based on these inputs, we transmitted these values into a Flask-based application, taking user inputs and implementing the model to predict future prices. Regarding the options, we used Flask to secure user input and used a python option pricing library to calculate the value of the option given the user’s constraints.

Challenges we ran into

The primary challenges we ran into were while training the linear regression model, as converting yahoo finance data to a pandas-readable data frame was challenging. Moreover, integrating Flask into our web app was hard, as we had to render user inputs to use the models. In essence, the primary challenges revolved around training the model and integrating the model with the front end.

Accomplishments that we're proud of

The primary accomplishment that we are proud of is the accuracy of the model. Our linear regression model has an accuracy of ninety-eight percent, which means that our predictions are very close to being accurate. This allows our product to be used by traders with a certain degree of confidence.

What we learned

Our learnings during this project were mainly around the yahoo finance library and Flask. The yahoo finance library was new to us, but this is very powerful as it can retrieve stock data and updates frequently as well. Moreover, integrating Flask with our linear regression model was another tool that we learned while creating this project.

What's next for Stock/Option Prediction

Regarding the future of this product, we hope to change the model from linear regression to a random forest regression model to achieve a higher level of accuracy with more concrete predictions. Beyond this, we can integrate more securities into the product, such as ETFs and Mutual Fund valuations, so we can have a higher user base and more capabilities of this application.

Built With

Share this project: