Abstract: Electricity demand forecasting plays an important role in short-term load allocation and long-term planning for future generation facilities and transmission augmentation. Planners must adopt a probabilistic view of potential peak demand levels, therefore density forecasts (providing estimates of the full probability distributions of the possible future values of the demand) are more helpful than point forecasts, and are necessary for utilities to evaluate and hedge the financial risk accrued by demand variability and forecasting uncertainty.

Electricity demand in a given season is subject to a range of uncertainties, including underlying population growth, changing technology, economic conditions, prevailing weather conditions (and the timing of those conditions), as well as the general randomness inherent in individual usage. It is also subject to some known calendar effects due to the time of day, day of week, time of year, and public holidays.

I will describe a comprehensive forecasting solution designed to take all the available information into account, and to provide forecast distributions from a few hours ahead to a few decades ahead. We use semi-parametric additive models to estimate the relationships between demand and the covariates, including temperatures, calendar effects and some demographic and economic variables. Then we forecast the demand distributions using a mixture of temperature simulation, assumed future economic scenarios, and residual bootstrapping. The temperature simulation is implemented through a new seasonal bootstrapping method with variable blocks.

The model is being used by the state energy market operators and some electricity supply companies to forecast the probability distribution of electricity demand in various regions of Australia. It also underpinned the Victorian Vision 2030 energy strategy.

Slides

• Adelaide Uni slides
• Later version given at Monash University, 16 May 2012.

Key papers

• Hyndman, R.J. and Fan, S. (2010) “Density forecasting for long-term peak electricity demand”, IEEE Transactions on Power Systems, 25(2), 1142–1153.
• Fan, S. and Hyndman, R.J. (2012) “Short-term load forecasting based on a semi-parametric additive model”. IEEE Transactions on Power Systems, 27(1), 134–141.

^aDepartment of Econometrics & Business Statistics, Monash University, Clayton, Victoria, Australia.
^bThe Australian Demographic & Social Research Institute, Australian National University, Canberra, ACT, Australia.

Demography, to appear.

Revised version: 20 April 2012.

Abstract:
When independence is assumed, forecasts of mortality for subpopulations are almost always divergent in the long term. We propose a method for coherent forecasting of mortality rates for two or more subpopulations, based on functional principal components models of simple and interpretable functions of rates. The product-ratio functional forecasting method models and forecasts the geometric mean of subpopulation rates and the ratio of subpopulation rates to product rates. Coherence is imposed by constraining the forecast ratio function through stationary time series models. The method is applied to sex-specific data for Sweden and state-specific data for Australia. Based on out-of-sample forecasts, the coherent forecasts are at least as accurate in overall terms as comparable independent forecasts, and forecast accuracy is homogenised across subpopulations.

Keywords: Mortality forecasting, coherent forecasts, functional data, Lee-Carter method, life expectancy, mortality, age pattern of mortality, sex-ratio.

Download working paper

Revised 10 January 2011

IEEE Transactions on Power Systems (2012), 27(1), 134–141.

Abstract
Short-term load forecasting is an essential instrument in power system planning, operation and control. Many operating decisions are based on load forecasts, such as dispatch scheduling of generating capacity, reliability analysis, and maintenance planning for the generators. Overestimation of electricity demand will cause a conservative operation, which leads to the start-up of too many units or excessive energy purchase, thereby supplying an unnecessary level of reserve. On the contrary, underestimation may result in a risky operation, with insufficient preparation of spinning reserve, causing the system to operate in a vulnerable region to the disturbance.

In this paper, semi-parametric additive models are proposed to estimate the relationships between demand and the driver variables. Specifically, the inputs for these models are calendar variables, lagged actual demand observations and historical and forecast temperature traces for one or more sites in the target power system. In addition to point forecasts, prediction intervals are also estimated using a modified bootstrap method suitable for the complex seasonality seen in electricity demand data. The proposed methodology has been used to forecast the half-hourly electricity demand for up to seven days ahead for power systems in the Australian National Electricity Market. The performance of the methodology is validated via out-of-sample experiments with real data from the power system, as well as through on-site implementation by the system operator.

Download working paper

Online published version

BMC Medical Research Methodology, 2012, to appear.

1. School of Public Health, La Trobe University, Bundoora, 3086Australia
2. School of Human Movement and Sport Sciences, University of Ballarat,Mount Helen, Victoria, 3353,Australia
3. Department of Econometrics and Business Statistics, Monash University, Clayton, 3800, Australia
4. VicHealth Centre for Tobacco Control, The Cancer Council Victoria,100 Drummond St, Carlton, Victoria, 3053
5. School of Public Health and Preventive Medicine, Monash University,Alfred Hospital, Melbourne, 3004,Australia.

ABSTRACT

Background: Chronic Obstructive Pulmonary Disease (COPD) is currently the fifth leading cause of death in Australia, and there are marked differences in mortality trends between men and women. In this study, we have sought to model and forecast age related changes in COPD mortality over time for men and women separately over the period 2006–2025.

Methods: Annual COPD death rates in Australia from 1922 to 2005 for age groups (50–54, 55–59, 60–64, 65–69, 70–74, 75–79, 80–84, 85+) were used. Functional time series models of age-specific COPD mortality rates for men and women were used, and forecasts of mortality rates were modelled separately for men and women.

Results: Functional time series models with four basis functions were fitted to each population separately. Twenty-year forecasts were computed, and indicated an overall decline. This decline may be slower for women than for men. By age, we expect similar rates of decline in men over time. In contrast, for women, forecasts for the age group 75–79 years suggest less of a decline over time compared to younger age groups.

Conclusions: By using a new method to predict age-specific trends in COPD mortality over time, this study provides important insights into at-risk age groups for men and women separately, which has implications for policy and program development.

Key words: COPD mortality, functional data analysis, tobacco consumption, forecasting

Online article

Journal of the American Statistical Association (2011) 106(496), 1513–1527.

Abstract
A new innovations state space modeling framework, incorporating Box-Cox transformations, Fourier series with time varying coefficients and ARMA error correction, is introduced for forecasting complex seasonal time series that cannot be handled using existing forecasting models. Such complex time series include time series with multiple seasonal periods, high frequency seasonality, non-integer seasonality and dual-calendar effects. Our new modelling framework provides an alternative to existing exponential smoothing models, and is shown to have many advantages. The methods for initialization and estimation, including likelihood evaluation, are presented, and analytical expressions for point forecasts and interval predictions under the assumption of Gaussian errors are derived, leading to a simple, comprehensible approach to forecasting complex seasonal time series. Our trigonometric formulation is also presented as a means of decomposing complex seasonal time series, which cannot be decomposed using any of the existing decomposition methods. The approach is useful in a broad range of applications, and we illustrate its versatility in three empirical studies where it demonstrates excellent forecasting performance over a range of prediction horizons. In addition, we show that our trigonometric decomposition leads to the identification and extraction of seasonal components, which are otherwise not apparent in the time series plot itself.

Keywords: exponential smoothing, Fourier series, prediction intervals, seasonality, state space models, time series decomposition.

Pre-publication working paper

Published paper

Data

• Call center data (day of week indicator)
• Gasoline data
• Turkish electricity data

I will look at the various facilities for time series forecasting available in R, concentrating on the forecast package. This package implements several automatic methods for forecasting time series including forecasts from ARIMA models, ARFIMA models and exponential smoothing models. I will also look more generally at how to go about forecasting non-seasonal data, seasonal data, seasonal data with high frequency, and seasonal data with multiple frequencies. Examples will be taken from my own consulting experience. I will give an overview of what’s possible and available and where it is useful, rather than give the mathematical details of any specific time series methods.

Slides
Examples

Abstract:

Electricity demand forecasting plays an important role in short-term load allocation and long-term planning for future generation facilities and transmission augmentation. Planners must adopt a probabilistic view of potential peak demand levels, therefore density forecasts (providing estimates of the full probability distributions of the possible future values of the demand) are more helpful than point forecasts, and are necessary for utilities to evaluate and hedge the financial risk accrued by demand variability and forecasting uncertainty.

Electricity demand in a given season is subject to a range of uncertainties, including underlying population growth, changing technology, economic conditions, prevailing weather conditions (and the timing of those conditions), as well as the general randomness inherent in individual usage. It is also subject to some known calendar effects due to the time of day, day of week, time of year, and public holidays.

I will describe a comprehensive forecasting solution designed to take all the available information into account, and to provide forecast distributions from a few hours ahead to a few decades ahead. We use semi-parametric additive models to estimate the relationships between demand and the covariates, including temperatures, calendar effects and some demographic and economic variables. Then we forecast the demand distributions using a mixture of temperature simulation, assumed future economic scenarios, and residual bootstrapping. The temperature simulation is implemented through a new seasonal bootstrapping method with variable blocks.

The model is being used by the state energy market operators and some electricity supply companies to forecast the probability distribution of electricity demand in various regions of Australia. It also underpinned the Victorian Vision 2030 energy strategy.

We evaluate the performance of the model by comparing the forecast distributions with the actual demand in some previous years. An important aspect of these evaluations is to find a way to measure the accuracy of density forecasts and extreme quantile forecasts.

Download slides

^a School of Geography and Environmental Science, Monash University, Melbourne, Australia
^b Department of Econometrics and Business Statistics, Monash University, Melbourne, Australia

Atmospheric Environment (2011), 45(1), 128–136.

Abstract
The influence of synoptic-scale circulations on air quality is an area of increasing interest to air quality management in regards to future climate change. This study presents an analysis where the dominant synoptic ‘types’ over the region of Melbourne, Australia are determined and linked to regional air quality. First, a self-organising map (SOM) is used to generate a time series of synoptic charts that classify the annual daily circulation affecting Melbourne into 20 different synoptic types. SOM results are then employed within the framework of a generalized additive model (GAM) to identify links between synoptic-scale circulations and observed changes air pollutant concentrations. The GAMs estimate shifts in pollutant concentrations under each synoptic type after controlling for long-term trends, seasonality, weekly emissions, spatial variation, and temporal persistence. Results showed the aggregate impact of synoptic circulations in the models to be quite modest as only 5.1% of the daily variance in O3, 4.7% in PM10, and 7.1% in NO2 were explained by shifts in synoptic circulations. Further analysis of the partial residual plots identified that despite a modest response at the aggregate level, individual synoptic categories had differential effects on air pollutants. In particular, increases of up to 40% in NO2 and PM10 and 30% in O3 occur when a synoptic conditions result in a north-easterly gradient wind over the Melbourne area. Additionally, NO2 and PM10 levels also showed increases of up to 40% when a strong high pressure system was centered directly over the Melbourne area. In sum, the unified approach of SOM and GAM proved to be a complementary suite of tools capable of identifying the entire range synoptic circulation patterns over a particular region and quantifying how they influence local air quality.

Keywords: air pollution, generalized additive models, self-organizing maps, and synoptic meteorology.

Working paper

1. Department of Econometrics & Business Statistics, Monash University, Clayton, Australia
2. The Australian Demographic & Social Research Institute, Australian National University, Canberra, Australia.

Demographic Research (2011), 25(5), 173–214.

Revised: 5 April 2011

Abstract:
Using the age– and sex-specific data of 14 developed countries, we compare the point and interval forecast accuracy and bias of ten principal component methods for forecasting mortality rates and life expectancy. The ten methods are variants and extensions of the Lee-Carter method. Based on one-step forecast errors, the weighted Hyndman-Ullah method provides the most accurate point forecasts of mortality rates and the Lee-Miller method is the least biased. For the accuracy and bias of life expectancy, the weighted Hyndman-Ullah method performs the best for female mortality and the Lee-Miller method for male mortality. While all methods underestimate variability in mortality rates, the more complex Hyndman-Ullah methods are more accurate than the simpler methods. The weighted Hyndman-Ullah method provides the most accurate interval forecasts for mortality rates, while the robust Hyndman-Ullah method provides the best interval forecast accuracy for life expectancy.

Keywords: Mortality forecasting, life expectancy forecasting, principal component methods, Lee-Carter method, interval forecasts, forecasting time series.

Online article

Analytical Chemistry (2011), 83(16), 6373–6380.

ABSTRACT: In industry as well as many areas of scientific research, data collected often contain a number of responses of interest for a chosen set of exploratory variables. Optimization of such multivariable multiresponse systems is a challenge well suited to genetic algorithms as global optimization tools. One such example is the optimization of coating surfaces with the required absolute and relative sensitivity for detecting analytes using devices such as sensor arrays. High-throughput synthesis and screening methods can be used to accelerate materials discovery and optimization; however, an important practical consideration for successful optimization of materials for arrays and other applications is the ability to generate adequate information from a minimum number of experiments. Here we present a case study to evaluate the efficiency of a novel evolutionary model-based multiresponse approach (EMMA) that enables the optimization of a coating while minimizing the number of experiments. EMMA plans the experiments and simultaneously models the material properties. We illustrate this novel procedure for materials optimization by testing the algorithm on a sol
gel synthetic route for production and optimization of a well studied amino-methyl-silane coating. The response variables of the coating have been optimized based on application criteria for micro– and macro-array surfaces. Spotting performance has been monitored using a fluorescent dye molecule for demonstration purposes and measured using a laser scanner. Optimization is achieved by exploring less than 2% of the possible experiments, resulting in identification of the most influential compositional variables. Use of EMMA to optimize control factors of a product or process is illustrated, and the proposed approach is shown to be a promising tool for simultaneously optimizing and modeling multivariable multiresponse systems.

Online paper