Inspiration

When you hear Machine Learning, the first use that comes to most peoples imaginations is using it to predict stock market trends in the Stock Market. Mirage is a step in that direction, as we try to identify four basic patterns: Bullish, Bearish, and Reversal patterns.

What it does

After giving a brief description of the patterns to the user, a camera appears on the screen. The user is then able to take a picture of a Stock Market chart, we recommend Trading View, then Mirage identifies any patterns that the chart exhibits.

How I built it

We gathered numerous photos of the different stock patterns, then we fed them into Apple's CoreML system. After that, we implemented our trained CoreML system into our app to use with our camera.

Challenges I ran into

Training an ML system to be accurate was a lot harder than we imagined. We never thought about how similar the photos needed to be, as outliers often can skew the data. The biggest challenge in the app was conquering the CoreML training.

Accomplishments that I'm proud of

We are proud that we trained a ML learning to have around a 60% accuracy! The app with the UI also came along nicely. Machine Learning is challenging, but not as intimidating as we originally imagined.

What I learned

I learned about CoreML, training a ML system, and implementing it into an app.

What's next for Mirage - A Technical Analysis Tool

A better trained ML, with more patterns! Ability to type in a stock ticker, then Mirage will analyze the tone of the most recent news articles for the stock, giving the user more information about news surrounding a stock.

Built With

Share this project:

Updates