blob: f7758ed812014162993bcbcf0e47723995666aa3 (
plain)
1
2
3
4
5
6
7
8
9
|
# lj_matrix
This is an exploration of a representation (or descriptor), based on the Lennard-Jones potential, for use in prediction of atomization energies (and possibly other properties) of molecules using Machine Learning (ML).
An implementation of existing representations, this new representation and most of the ML routine is created from the, mostly, ground up.
The dataset used is the QM7 dataset obtained from the [QML tutorial repository](https://github.com/qmlcode/tutorial). On the other hand, the periodic table of elements data was retrieved from [this handy Gist](https://gist.github.com/GoodmanSciences/c2dd862cd38f21b0ad36b8f96b4bf1ee).
*NOTE*: This is not supposed to be a python package (for now).
|
