We trained a machine learning model by using mixture density network (MDN) algorithm to quickly and efficiently predict the interior structure of rocky exoplanets with masses between 0.1 and 10 Earth masses.
Fork and clone a copy of the Rocky_Exoplanets
repository on to your local machine.
Download Anaconda
and install it on your machine.
Create a conda
environment called Rocky_Exoplanets
and install all the necessary dependencies:
$ conda create -n Rocky_Exoplanets pip python jupyter tensorflow=1.15
Activate the Rocky_Exoplanets
environment:
$ conda activate Rocky_Exoplanets
Change into your local copy of the Rocky_Exoplanets
repo:
$ cd /you own path/Rocky_Exoplanets
Install the requirments for predicing in the current Conda environment:
$ pip install -r requirements.txt
Open Jupyter Notebook
and load the file MDN_prediction.ipynb
:
$ jupyter notebook
At this point you are ready to start investigating the interiors of rocky exoplanets!