Star System Generator

Demo:

HOW TO RUN:

Clone the repo and open the project directory.

git clone https://github.com/leiDnedyA/interactive-exoplanet-predictor
cd interactive-exoplanet-predictor
./setup # installs dependencies for API and frontend

Run in development mode

1. Run ./devstart to start the API and frontend. The API and client should be started automatically, and your browser will open an instance of the client.

Build and run

1. run ./buildstart
2. Open your browser to http://localhost:3000/. If you've changed any of the ports in .env, you may have to open localhost with at different port.

What happens in the model:

We implemented the Random Forest machine learning model mainly because it has the highest accuracy of all the models we tested. We used the scikit-learn library to implement the model. We used the data from the Exoplanet Archive to train the model. We used the following features to train the model:

• Stellar Mass
• Stellar Radius
• Stellar Density
• Stellar Temperature
• Stellar Metallicity
• Stellar Radial Velocity
• Stellar Surface Gravity
• Stellar Age

By training and testing the model with a 0.2 sample rate, we got a 92% (First Release) / 98% (Current Beta Version) accuracy score, which is expected to rise each monthly data update. In order to predict the number of planets a star could have, we used sklearn's RandomForest model.

We are currently working on:

• Allowing multiple predictions to be made at once as a big data set.
• Adding more features to the model to increase accuracy.
• Updating the model with new data each month. (Completed)
• Comparing the accuracy of the model to other machine learning models