ds-wook / power-usage-predict

🥉2023 Power Consumption Prediction AI Competition🥉

power-usage-predict

2023 Power Consumption Prediction AI Competition.
Developing an AI model that predicts the power consumption at a specific point in time using building information and spatiotemporal data.

Setting

• CPU: i7-11799K core 8
• RAM: 32GB
• GPU: NVIDIA GeForce RTX 3090 Ti

Cross Validation

• Stratified Group KFold

Ensemble Strategy

• Ensemble Folds with Median

Model Process

• Boosting is All you need

Requirements

By default, hydra-core==1.1.0 was added to the requirements given by the competition. For pytorch, refer to the link at https://pytorch.org/get-started/previous-versions/ and reinstall it with the right version of pytorch for your environment.

You can install a library where you can run the file by typing:

$ conda env create --file environment.yaml

Run code

Code execution for the new model is as follows:

Running the learning code shell.

$ sh scripts/run.sh

Examples are as follows.

python src/clustering.py

for model in xgboost lightgbm catboost; do
 python src/train.py models=$model
 python src/predict.py models=$model
done

python src/teach.py
python src/ensemble.py

Benchmark

XGBoost-custom-loss: 5.5316 LightGBM-tweedie-loss: 5.8699 Categorical-Non-Catboost: 5.5252 Categorical-Catboost: 5.7216

The NN model has a significant performance difference compared to the boosting. Ensemble results also appeared to have a greater impact than other models.

Submit

Doesn't Work

• meta feature: mean features
• forcasting model: NBeat is not performance

Reference

• Model
• Loss
• Ensemble