analokmaus/idao2020-online

All you need is linear regression

IDAO 2020 online roud

Team: QuMantumPhysicists

Directory

.
├── input/
  ├── train.csv
  └── Track 1
    └── test.csv
├── requirements.txt
├── Makefile
├── README.md
├── model_description.PNG
├── fit_param_dictAA.dump
├── fit_param_dictB2.dump 
├── lr_hyperparam.csv
├── main.sh
├── trackA_predict.py
├── trackB_predict.py
└── train.py

• common
  • input/
    • input data folder from https://disk.yandex.ru/d/0zYx00gSraxZ3w
  • requirements.txt
    • development environment
  • train.py
    • fit by linear regression and save results
    • generate fit_param_dictAA.dump and fit_param_dictB2.dump
  • lr_hyperparam.csv
    • list of hyper parameters
• for track A
  • fit_param_dictAA.dump
  • trackA_predict.py
    • make prediction
• for track B
  • fit_param_dictB2.dump
  • trackB_predict.py
    • make prediction
  • Makefile
  • main.sh

---

Instructions

1. Place input data in input/ (shown in Directory section).
2. If necessary, install packages in requirements.txt.
3. Train models for both tracks: python train.py.
   • Pretrained models(fit_param_dictAA.dump and fit_param_dictB2.dump) will be generated.
   • It takes around 1 hour with 4 cores.

for track A:

4. Make submission: python trackA_predict.py.
   • submission_trackA.csv will be generated.

for track B:

4. Zip current working directory.

---

Model description

All you need is linear regression