Skip to content

CozyDoomer/Predicting-Molecular-Properties

Folders and files

NameName
Last commit message
Last commit date

Latest commit

4e85303 · Jan 7, 2022

History

50 Commits
Aug 29, 2019
Jul 21, 2019
Dec 22, 2019
Jan 7, 2022
Jul 20, 2019
Aug 29, 2019
Jul 30, 2019
Jul 20, 2019
Jul 20, 2019
Aug 29, 2019
Aug 29, 2019
Aug 29, 2019
Aug 29, 2019
Jul 20, 2019

Repository files navigation

Predicting Molecular Properties

Kaggle competition link

Solution writeup

In this competition, you will develop an algorithm that can predict the magnetic interaction between two atoms in a molecule (i.e., the scalar coupling constant).

Dataformat Metric Prediction
structured data (graph based) log mean average error regression

Started working on this competition using lightgbm and then used a modified implementation of a message passing neural network.

Notes

A lot of additional data that is not usable directly because it's not contained in the test set.

Domain knowledge about atom interaction in molecules was really important (to a certain degree).

Most of the features were calculated using rdkit and openbabel.

local validation for message passing neural network

Per coupling type:

  • 1JHC: -1.371
  • 2JHC: -2.229
  • 3JHC: -1.975
  • 1JHN: -1.538
  • 2JHN: -2.504
  • 3JHN: -2.517
  • 2JHH: -2.501
  • 3JHH: -2.383

average local log mae: -2.12

Placement

top 2%

leaderboard score placement
public -2.37190 43/2757
private -2.36477 42/2757