MSBD5001IndividualProject

This repo contains the solution produced by Wong, Tsz Fung for the MSBD5001 Individual Project on kaggle: https://www.kaggle.com/c/msbd5001-fall2020/overview

Team name: Thomas Wong

Public Leaderboard score: 9.93795 & 10.03450.

Private Leadboard score: 10.64356 (Ranked 7th excluding deleted accounts due to people cheating)

2 submissions were chosen as final.

You can find the submission csv files under 5001/.

Description

The local CV score is around 9.75 MSE.

This CV score is obtained by taking the mean of 100 instances of LightGBM model with CV. (Code not provided)

Running Locally

The python version I used is Python 3.8.

Install the required packages by running:

pip install -r requirements.txt

Running on Colab (Optional)

The notebook was originally run on Google Colab.

But running on Colab requires you to install the holidays package by running:

!pip install holidays

How to reproduce the results

Simply run the cells in the jupyter notebook. It should generate a submission file that scores 10.03450 on the public leaderboard. The lower score was obtained when the random_state of CV was removed.