자율주행 센서의 안테나 성능 예측 AI 경진대회

Public score 2th 1.89526 | Private score 2th 1.91352

• 주최 : LG AI Research
• 주관 : DACON
• https://dacon.io/competitions/official/235927/overview/description

Usage

• train.sh : If the trained model not exist, reproduce the model.
• inference.sh : If the trained model exist, only inference without training.

Pipeline

Setting Environment

• python version >= 3.6

Make virtual env

$ python3 -m venv pyenv
$ source ./pyenv/bin/activate

Install requirements

$ (pyenv) pip install --upgrade pip
$ (pyenv) pip install -r requirements.txt 

Train Run Shell

$ (pyenv) sh ./train.sh

or

$ (pyenv) python ./src/preprocess.py
$ (pyenv) python ./src/train.py

Inference Run Shell

$ (pyenv) sh ./inference.sh

or

$ (pyenv) python ./src/preprocess.py
$ (pyenv) python ./src/inference.py

File Description

feature.py : feature engineering class py
model.py : model class py
preprocess.py : preprocess activate code
train.py : train activate code
utils.py : utils func py
inference.py : inference activate code

Requirements

numpy
pandas
tqdm
scikit-learn
lightgbm==3.3.2
xgboost==1.6.1
catboost==1.0.6

Directory Structure

/workspace
├── model
│   ├── clean
│   │   ├── level0
│   │   ├── level1
│   ├── noise
│   │   ├── level0
│   │   ├── level1
├── open
│   ├── meta
│   │   ├── sample_submission.csv
│   │   ├── test.csv
│   │   ├── train.csv
├── output
│   ├── clean
│   │   ├── submission.csv
│   ├── final
│   │   ├── submission.csv
│   ├── noise
│   │   ├── submission.csv
├── refine
│   ├── clean
│   │   ├── raw
│   │   ├── scale
│   ├── noise
│   │   ├── raw
│   │   ├── scale
├── src
│   ├── feature.py
│   ├── inference.py
│   ├── model.py
│   ├── preprocess.py
│   ├── train.py
│   ├── utils.py
├── inference.sh
├── train.sh
      .
      .
      .