EnergyConsumptionCase

Description

Predicting energy consumption using an LSTM Autoencoder and XGBoost.

Goal

As a DS student, my goals with this project are only to study a little bit more about time series cases, the implementation of an LSTM Autoencoder and a XGBoost model for prediction.

Dataset

https://www.kaggle.com/datasets/robikscube/hourly-energy-consumption

Folder Structure

The project has been organized in a linear manner to facilitate navigation and understanding of the step-by-step process.

├── automations
│   │   ├── folders_creation.py
├── data
│   ├── d01_raw
│   │   ├── PJME_hourly.csv
│   ├── d02_intermediate
│   │   ├── pjme_n00.csv
│   │   ├── pjme_n01.csv
│   │   ├── pjme_test.csv
│   │   ├── pjme_train.csv
├── model
│   ├── xgboost_gs.pkl
│   ├── xgboost_params.txt
├── notebooks
│   ├── n00_data_preparation.ipynb
│   ├── n01_exploratory_data_analysis.ipynb
│   ├── n02_data_split.ipynb
│   ├── n03_feature_engineering.ipynb
│   ├── n04_xgboost_training_validation.ipynb
│   ├── n04_lstm_training_validation.ipynb
└── README.md

https://towardsdatascience.com/step-by-step-understanding-lstm-autoencoder-layers-ffab055b6352 https://towardsdatascience.com/lstm-autoencoder-for-extreme-rare-event-classification-in-keras-ce209a224cfb https://machinelearningmastery.com/lstm-autoencoders/

Linkedin

https://www.linkedin.com/in/sabrina-otoni-da-silva-22525519b/