Exo_Predict

This machine learning planet classification package allows users to predict the number of exoplanets a user is likely to find in a given exoplanet system. It does this based solely on the stellar mass, radius, temperature, and the exoplanet discovery method employed.

Installation

To install this package, run the following from the command line:

pip install exo_predict

Then, you can import the package in python:

import exo_predict

Functionality

Using the default exo_predict module, you can use our model trained on all confirmed planets from the NASA Exoplanet Archive (https://exoplanetarchive.ipac.caltech.edu/) in order to predict whether or not the systems in your data contain one or multiple planets.

If you'd rather use your own data train a model, you can feel free to do so with our exo_model_predict module. It takes in your training data to create a model, then hands you back the model and a testing dataset. This module also allows you to input new data to try out once you've tried out your model with the predict_exoplanets function.

Tutorials

Two tutorials are included with some sample data so you can try this out for yourself! Find them in the Tutorials directory.

Requirements

You will need the following to use this package:

numpy
pandas
sklearn
xgboost
matplotlib
pickle