Follow-up forecasting forum in Eindhoven

Last October I gave a 3-day masterclass on “Forecasting with R” in Eindhoven, Netherlands. There is a follow-up event planned for Tuesday 18 April 2017. It is particularly designed for people who attended the 3-day class, but if anyone else wants to attend they would be welcome.

Please register here if you want to attend. Continue reading →

Data Science for Managers (short course)

I am teaching part of a short-course on Data Science for Managers from 10-12 October in Melbourne.

Course Overview

The impact of Data Science on modern business is second only to the introduction of computers. And yet, for many businesses the barrier of entry remains too high due to lack of knowhow, organisational inertia, difficulties in hiring the right manpower, an apparent need for upfront commitment, and more.

This course is designed to address these barriers, giving the necessary knowledge and skills to flesh out and manage Data Science functions within your organisation, taking the anxiety-factor out of the Big Data revolution and demonstrating how data-driven decision-making can be integrated into one’s organisation to harness existing advantages and to create new opportunities.

Assuming minimal prior knowledge, this course provides complete coverage of the key aspects, including data wrangling, modelling and analysis, predictive-, descriptive- and prescriptive-analytics, data management and curation, standards for data storage and analysis, the use of structured, semi-structured and unstructured data as well as of open public data, and the data-analytic value chain, all covered at a fundamental level.

More details available at

Early-bird bookings close in a few days.


The bias-variance decomposition

This week, I am teaching my Business Analytics class about the bias-variance trade-off. For some reason, the proof is not contained in either ESL or ISL, even though it is quite simple. I also discovered that the proof currently provided on Wikipedia makes little sense in places.

So I wrote my own for the class. It is longer than necessary to ensure there are no jumps that might confuse students.
Continue reading →

Resources for the FPP book

The FPP resources page has recently been updated with several new additions including

  • R code for all examples in the book. This was already available within each chapter, but the examples have been collected into one file per chapter to save copying and pasting the various code fragments.
  • Slides from a course on Predictive Analytics from the University of Sydney.
  • Slides from a course on Economic Forecasting from the University of Hawaii.

If any one using the book has other material that could be made available, please send them to me. For example, recorded lectures, slides, additional examples, assignments, exam questions, solutions, etc.

Forecasting with R in WA

On 23-25 September, I will be running a 3-day workshop in Perth on “Forecasting: principles and practice” mostly based on my book of the same name.

Workshop participants will be assumed to be familiar with basic statistical tools such as multiple regression, but no knowledge of time series or forecasting will be assumed. Some prior experience in R is highly desirable.

Venue: The University Club, University of Western Australia, Nedlands WA.

Day 1:
Forecasting tools, seasonality and trends, exponential smoothing.
Day 2:
State space models, stationarity, transformations, differencing, ARIMA models.
Day 3:
Time series cross-validation, dynamic regression, hierarchical forecasting, nonlinear models.

The course will involve a mixture of lectures and practical sessions using R. Each participant must bring their own laptop with R installed, along with the fpp package and its dependencies.

For costs and enrolment details, go to

Student forecasting awards from the IIF

At the IIF annual board meeting last month in Rotterdam, I suggested that we provide awards to the top students studying forecasting at university level around the world, to the tune of $100 plus IIF membership for a year. I’m delighted that the idea met with enthusiasm, and that the awards are now available. Even better, my second year forecasting subject has been approved for an award.

The IIF have agreed to fund awards for 20 forecasting courses to start with. I believe they have already had several applications, so any other forecasting lecturers out there will need to be quick if they want to be part of it.

Creating a handout from beamer slides

I’m about to head off on a speaking tour to Europe (more on that in another post) and one of my hosts has asked for my powerpoint slides so they can print them. They have made two false assumptions: (1) that I use powerpoint; (2) that my slides are static so they can be printed.

Instead, I produced a cut-down version of my beamer slides, leaving out some of the animations and other features that will not print easily. Then I produced a pdf file with several slides per page. Continue reading →

Interpreting noise

04_03_2_prevWhen watching the TV news, or reading newspaper commentary, I am frequently amazed at the attempts people make to interpret random noise.

For example, the latest tiny fluctuation in the share price of a major company is attributed to the CEO being ill. When the exchange rate goes up, the TV finance commentator confidently announces that it is a reaction to Chinese building contracts. No one ever says “The unemployment rate has dropped by 0.1% for no apparent reason.”

What is going on here is that the commentators are assuming we live in a noise-free world. They imagine that everything is explicable, you just have to find the explanation. However, the world is noisy — real data are subject to random fluctuations, and are often also measured inaccurately. So to interpret every little fluctuation is silly and misleading. Continue reading →

Fast computation of cross-validation in linear models

The leave-one-out cross-validation statistic is given by
\text{CV} = \frac{1}{N} \sum_{i=1}^N e_{[i]}^2,

where ${e_{[i]} = y_{i} – \hat{y}_{[i]}} $, the observations are given by $y_{1},\dots,y_{N}$, and $\hat{y}_{[i]}$ is the predicted value obtained when the model is estimated with the $i\text{th}$ case deleted. This is also sometimes known as the PRESS (Prediction Residual Sum of Squares) statistic.

It turns out that for linear models, we do not actually have to estimate the model $N$ times, once for each omitted case. Instead, CV can be computed after estimating the model once on the complete data set. Continue reading →