M10 - Advanced Forecasting Models

Learning Objectives

The learning goals for this module are:

• Discuss challenges in forecasting higher frequency time series;
• Discuss complex seasonality;
• Go over some functions in R that can handle multiple seasonal pattern;
• Develop a code for forecasting daily electricity demand.

Resources

This module is a collection of models and R functions for forescasting more granular data with multiple seasonality. Please refer to the Rmd file available here.

Additional reading to support the models introduced in this module are provided in the Rmd file. The data file used in this module is too large to be uploaded to Github, you may directly download the data household_power_consumption.zip using this link.

• Forecasting: Principles and Practice by Rob J Hyndman and George Athanasopoulos, Section 11: Advanced Forecasting Methods

Recordings

This module is explained in four videos that go over a Rmd file that contains code used to forecast higher frequency time series and discussion regarding complex seasonalities and how to model them with R.

• Importing and Preparing the Data Set
  

• Decomposing Multiple Seasonal Series
  

• Forecasting - Part 1
  

• Forecasting - Part 2
  

Deliverables

For this module you will complete Assignment 8 - Forecasting Competition. Please refer to the Assignments tab for due dates.