Time series forecasting models from small dataset. Project assignment for the Machine Learining classes

Piotr Kőpplinger​, Ewa Siedlarczyk​, Piotr Urbańczyk

Overview

The goal of this project is to evaluate the performance of several time series forecasting models on a small biological dataset. The dataset captures microbiome samples and their dynamics over time. Our evaluation focuses on two main metrics:

• Mean Absolute Scaled Error (MASE): A scale-independent measure of prediction accuracy.
• Computation Time: The time required to train and forecast using each model.

How to Run

1. Clone the Repository:

   https://github.com/ewa-siedlarczyk/UM-TimeSeries.git
   cd UM-TimeSeries

2. Create a virtual environment and install the dependencies

   python -m venv venv
   source venv/bin/activate  # `venv\Scripts\activate` on windows
   pip install -r requirements.txt

3. Run jupyter notebook

Models tested

• ARIMA
• ETS
• RNN
• LSTM
• TCN
• TimesNet
• TFT