Time-series-DS-project

This is a project for the Data Science Course in DCI Masters at the universty of Paris Cité for the year 2023-2024.

This projects evaluates time series forecasting methods through participation in two private Kaggle competitions.

In summary, we tried the following models: (some were more explored than others)

• ARIMA
• XGBOOST
• LSTM
• Costume Residual-CNN-LSTM network
• Temporal Convolutional Network

Getting Started

• Python 3.9.0

• install packages by running "pip install -r requirements.txt"

  note: Some deep learning models require extra packages like Pytorch.