Time Series Analysis and User Behaviour Learning to forecast power usage in a household.
The Dataset used to test the performance of various models for the task at hand are GreenD and UkDale. The resampled files of these datasets are present in the Data folder. The models can be used for any new data, provided they are available in the given format.
- Clone the repository
git clone https://github.com/Home-Electricity-Manager/Power_Forecasting.git
- Install the dependencies
pip install --r requirements.txt
- Preprocess the data to equipment-wise usable format
python Data/preprocess.py --src_dir <path-to-directory-containing-data> --src_file_name <common-prefix-for-all-equipments> --tar_dir <target-directory-to-place-preprocessed-data> --num_equips <number-of-equipments> --n_steps <number-of-timesteps, default=128>
- ARIMA
python Models/ARIMA.py --src_dir <path-to-directory-containing-preprocessed-data> --tar_dir <path-where-the-outputs-need-to-be-placed> --eq_num <equipment-number-to-be-tested> --num_preds <number-of-predictions-to-be-made> --draw_graphs <bool-variable--whether-to-draw-graphs-or-not>
- CNN_LSTM
python Models/CNN_LSTM.py --src_dir <path-to-directory-containing-preprocessed-data> --tgt_dir <path-where-the-outputs-need-to-be-placed> --eq_num <equipment-number-to-be-tested> --n_test <number-of-predictions-to-be-made> --n_models <numer-of-CNN_LSTM-models-to-be-made> --draw_graphs <bool-variable--whether-to-draw-graphs-or-not>
- Ensemble
python Models/Ensemble.py --src_dir <path-to-directory-containing-preprocessed-data> --tgt_dir <path-where-the-outputs-need-to-be-placed> --eq_num <equipment-number-to-be-tested> --n_test <number-of-predictions-to-be-made> --n_models <numer-of-CNN_LSTM-models-to-be-made> --draw_graphs <bool-variable--whether-to-draw-graphs-or-not>
python Models/StartFinish.py --data_path <path-to-directory-containing-data-to-apply-on> --target_path <path-where-the-outputs-need-to-be-placed> --num_preds <number-of-predictions-to-be-worked-on>