Time Series Forecast for Ozone Level: Applications of Exponential Smoothing & ARIMA models

Project Summary

• The project is to apply the two most widely-used approaches, i.e. Exponential Smoothing & ARIMA, for time series forecast.
• Ozone dataset is the subject of the experiment. The dataset is obtained from UCI Machine Learning Repository.
• Theoretical concepts of the two approaches are discussed so that approriate versions of Exponential Smoothings & ARIMAs are logically selected for the given dataset.
• Protocol of time series forecast is proposed. It includes:
  • EDA
  • Time Series Decomposition
  • Auto-Correlation Assessment
  • Applications of appropriate versions of Exponential Smoothings & ARIMAs
  • Cross validation of forecasts to calculate models' mean-squared errors
  • Choosing the best models.
• The details of the project is documented in the pdf file and the webpage.

Technology

R

Getting Started

1. Clone this repository (for help see this tutorial).
2. Run the code.R file in the data_n_code folder by using RStudio.
3. Follow along the comments in the the code.R file.

License

This project is under the MIT license.