Short-term electricity demand and gas price forecasts using wavelet transforms and adaptive models

This paper presents some forecasting techniques for energy demand and price prediction, one day ahead. These techniques combine wavelet transform (WT) with fixed and adaptive machine learning/time series models (multi-layer perceptron (MLP), radial basis functions, linear regression, or GARCH). To create an adaptive model, we use an extended Kalman filter or particle filter to update the parameters continuously on the test set. The adaptive GARCH model is a new contribution, broadening the applicability of GARCH methods. We empirically compared two approaches of combining the WT with prediction models: multicomponent forecasts and direct forecasts. These techniques are applied to large sets of real data (both stationary and non-stationary) from the UK energy markets, so as to provide comparative results that are statistically stronger than those previously reported. The results showed that the forecasting accuracy is significantly improved by using the WT and adaptive models. The best models on the electricity demand/gas price forecast are the adaptive MLP/GARCH with the multicomponent forecast; their MSEs are 0.02314 and 0.15384 respectively.

Density forecasts of crude-oil prices using option-implied and ARCH-type models

The predictive accuracy of competing crude-oil price forecast densities is investigated for the 1994–2006 period. Moving beyond standard ARCH type models that rely exclusively on past returns, we examine the benefits of utilizing the forward-looking information that is embedded in the prices of derivative contracts. Risk-neutral densities, obtained from panels of crude-oil option prices, are adjusted to reflect real-world risks using either a parametric or a non-parametric calibration approach. The relative performance of the models is evaluated for the entire support of the density, as well as for regions and intervals that are of special interest for the economic agent. We find that non-parametric adjustments of risk-neutral density forecasts perform significantly better than their parametric counterparts. Goodness-of-fit tests and out-of-sample likelihood comparisons favor forecast densities obtained by option prices and non-parametric calibration methods over those constructed using historical returns and simulated ARCH processes. © 2010 Wiley Periodicals, Inc. Jrl Fut Mark 31:727–754, 2011

Wind power forecasts using Gaussian processes and numerical weather prediction

Since wind at the earth's surface has an intrinsically complex and stochastic nature, accurate wind power forecasts are necessary for the safe and economic use of wind energy. In this paper, we investigated a combination of numeric and probabilistic models: a Gaussian process (GP) combined with a numerical weather prediction (NWP) model was applied to wind-power forecasting up to one day ahead. First, the wind-speed data from NWP was corrected by a GP, then, as there is always a defined limit on power generated in a wind turbine due to the turbine controlling strategy, wind power forecasts were realized by modeling the relationship between the corrected wind speed and power output using a censored GP. To validate the proposed approach, three real-world datasets were used for model training and testing. The empirical results were compared with several classical wind forecast models, and based on the mean absolute error (MAE), the proposed model provides around 9% to 14% improvement in forecasting accuracy compared to an artificial neural network (ANN) model, and nearly 17% improvement on a third dataset which is from a newly-built wind farm for which there is a limited amount of training data. © 2013 IEEE.

Three essays on corporate governance and meeting-beating or missing analyst forecasts

The beginning of the 21st century was plagued with unprecedented instances of corporate fraud. In an attempt to address apparent non-existent or “broken” corporate governance policies, sweeping measures of financial reporting reform ensued, having specific requirements relating to the composition of audit committees, the interaction between audit committees and external auditors, and procedures concerning auditors’ assessment of client risk. The purpose of my dissertation is to advance knowledge about “good” corporate governance by examining the association between meeting-or-beating analyst forecasts and audit fees, audit committee compensation, and audit committee tenure and “busyness”. Using regression analysis, I found the following: (1) the frequency of meeting-or-just beating (just missing) analyst forecasts is negatively (positively) associated with audit fees, (2) the extent by which a firm exceeds analysts’ forecasts is positively (negatively) associated with audit committee compensation that is predominately equity-based (cash-based), and (3) the likelihood of repeatedly meeting-or-just beating analyst forecasts is positively associated with audit committee tenure and “busyness”. These results suggest that auditors consider clients who frequently meet-or-just beat forecasts as being less “risky”, and clients that frequently just miss as being more “risky”. The results also imply that cash-based director compensation is more successful in preserving the effectiveness of the audit committee’s financial reporting oversight role, that equity-based compensation motivates independent audit committee directors to focus on short-term performance thereby aligning their interests with management, and that audit committee director tenure and the degree of director “busyness” can affect an audit committee member’s effectiveness in providing financial reporting oversight. Collectively, my dissertation provides additional insights regarding corporate governance practices and informs policy-makers for future relevant decisions.^

Monitoring the Accuracy of Multiple Occupancy Forecasts

Corporate executives closely monitor the accuracy of their hotels' occupancy fore- casts since important decisions are based upon these predictions. This study lists the criteria for selecting an appropriate error measure. It discusses several evaluation methods focusing on statistical significance tests and demonstrates the use of two adequate evaluation methods: Mincer- Zamowitz's efficiency test and Wilcoxon's Non-Parametric Matched-Pairs Signed- Ranks test.

Essays in book -tax income differences and analyst forecasts /recommendations

Prior research suggests that book-tax income differences (BTD) relate to both firms' earnings quality and operating performance. In this dissertation, I explore whether and how financial analysts signal the implications of BTD efficiently. This dissertation is comprised of three essays on BTD. The three essays seek to develop a better understanding of how financial analysts utilize information reflected in BTD (derived from the ratio of taxable income to book income). The first essay is a review and discussion of prior research regarding BTD. The second essay of this dissertation investigates the role of BTD in indicating the consensus and dispersion of analyst recommendations. I find that sell recommendations are positively related to BTD. I also document that analyst coverage has a positive effect on the standard deviation of consensus recommendations with respect to BTD. The third essay is an empirical analysis of analysts' forecast optimism, analyst coverage, and BTD. I find a negative association between forecast optimism and BTD. My results are consistent with a larger BTD being associated with less forecast bias. Overall, I interpret the sum of the evidence as being consistent with BTD reflecting information about earnings quality, and consistent with analysts examining and using this information in making decisions regarding both forecasts and recommendations.

Deterministic and probabilistic verification of multi-model meteorological forecasts on the subseasonal timescale

In questo studio, un multi-model ensemble è stato implementato e verificato, seguendo una delle priorità di ricerca del Subseasonal to Seasonal Prediction Project (S2S). Una regressione lineare è stata applicata ad un insieme di previsioni di ensemble su date passate, prodotte dai centri di previsione mensile del CNR-ISAC e ECMWF-IFS. Ognuna di queste contiene un membro di controllo e quattro elementi perturbati. Le variabili scelte per l'analisi sono l'altezza geopotenziale a 500 hPa, la temperatura a 850 hPa e la temperatura a 2 metri, la griglia spaziale ha risoluzione 1 ◦ × 1 ◦ lat-lon e sono stati utilizzati gli inverni dal 1990 al 2010. Le rianalisi di ERA-Interim sono utilizzate sia per realizzare la regressione, sia nella validazione dei risultati, mediante stimatori nonprobabilistici come lo scarto quadratico medio (RMSE) e la correlazione delle anomalie. Successivamente, tecniche di Model Output Statistics (MOS) e Direct Model Output (DMO) sono applicate al multi-model ensemble per ottenere previsioni probabilistiche per la media settimanale delle anomalie di temperatura a 2 metri. I metodi MOS utilizzati sono la regressione logistica e la regressione Gaussiana non-omogenea, mentre quelli DMO sono il democratic voting e il Tukey plotting position. Queste tecniche sono applicate anche ai singoli modelli in modo da effettuare confronti basati su stimatori probabilistici, come il ranked probability skill score, il discrete ranked probability skill score e il reliability diagram. Entrambe le tipologie di stimatori mostrano come il multi-model abbia migliori performance rispetto ai singoli modelli. Inoltre, i valori più alti di stimatori probabilistici sono ottenuti usando una regressione logistica sulla sola media di ensemble. Applicando la regressione a dataset di dimensione ridotta, abbiamo realizzato una curva di apprendimento che mostra come un aumento del numero di date nella fase di addestramento non produrrebbe ulteriori miglioramenti.

Eliciting and Aggregating Forecasts When Information is Shared

Using the wisdom of crowds---combining many individual forecasts to obtain an aggregate estimate---can be an effective technique for improving forecast accuracy. When individual forecasts are drawn from independent and identical information sources, a simple average provides the optimal crowd forecast. However, correlated forecast errors greatly limit the ability of the wisdom of crowds to recover the truth. In practice, this dependence often emerges because information is shared: forecasters may to a large extent draw on the same data when formulating their responses.

To address this problem, I propose an elicitation procedure in which each respondent is asked to provide both their own best forecast and a guess of the average forecast that will be given by all other respondents. I study optimal responses in a stylized information setting and develop an aggregation method, called pivoting, which separates individual forecasts into shared and private information and then recombines these results in the optimal manner. I develop a tailored pivoting procedure for each of three information models, and introduce a simple and robust variant that outperforms the simple average across a variety of settings.

In three experiments, I investigate the method and the accuracy of the crowd forecasts. In the first study, I vary the shared and private information in a controlled environment, while the latter two studies examine forecasts in real-world contexts. Overall, the data suggest that a simple minimal pivoting procedure provides an effective aggregation technique that can significantly outperform the crowd average.

The efficiency of multivariate macroeconomic forecasts

We examine the efficiency of multivariate macroeconomic forecasts by estimating a vector autoregressive model on the forecast revisions of four variables (GDP, inflation, unemployment and wages). Using a data set of professional forecasts for the G7 countries, we find evidence of cross‐series revision dynamics. Specifically, forecasts revisions are conditionally correlated to the lagged forecast revisions of other macroeconomic variables, and the sign of the correlation is as predicted by conventional economic theory. This indicates that forecasters are slow to incorporate news across variables. We show that this finding can be explained by forecast underreaction.

Investment decision-making under risk (reliability) and uncertainty for infrastructure asset management

Risks and uncertainties are inevitable in engineering projects and infrastructure investments. Decisions about investment in infrastructure such as for maintenance, rehabilitation and construction works can pose risks, and may generate significant impacts on social, cultural, environmental and other related issues. This report presents the results of a literature review of current practice in identifying, quantifying and managing risks and predicting impacts as part of the planning and assessment process for infrastructure investment proposals. In assessing proposals for investment in infrastructure, it is necessary to consider social, cultural and environmental risks and impacts to the overall community, as well as financial risks to the investor. The report defines and explains the concept of risk and uncertainty, and describes the three main methodology approaches to the analysis of risk and uncertainty in investment planning for infrastructure, viz examining a range of scenarios or options, sensitivity analysis, and a statistical probability approach, listed here in order of increasing merit and complexity. Forecasts of costs, benefits and community impacts of infrastructure are recognised as central aspects of developing and assessing investment proposals. Increasingly complex modelling techniques are being used for investment evaluation. The literature review identified forecasting errors as the major cause of risk. The report contains a summary of the broad nature of decision-making tools used by governments and other organisations in Australia, New Zealand, Europe and North America, and shows their overall approach to risk assessment in assessing public infrastructure proposals. While there are established techniques to quantify financial and economic risks, quantification is far less developed for political, social and environmental risks and impacts. The report contains a summary of the broad nature of decision-making tools used by governments and other organisations in Australia, New Zealand, Europe and North America, and shows their overall approach to risk assessment in assessing public infrastructure proposals. While there are established techniques to quantify financial and economic risks, quantification is far less developed for political, social and environmental risks and impacts. For risks that cannot be readily quantified, assessment techniques commonly include classification or rating systems for likelihood and consequence. The report outlines the system used by the Australian Defence Organisation and in the Australian Standard on risk management. After each risk is identified and quantified or rated, consideration can be given to reducing the risk, and managing any remaining risk as part of the scope of the project. The literature review identified use of risk mapping techniques by a North American chemical company and by the Australian Defence Organisation. This literature review has enabled a risk assessment strategy to be developed, and will underpin an examination of the feasibility of developing a risk assessment capability using a probability approach.

Time series analysis of a Web search engine transaction log

In this paper, we use time series analysis to evaluate predictive scenarios using search engine transactional logs. Our goal is to develop models for the analysis of searchers’ behaviors over time and investigate if time series analysis is a valid method for predicting relationships between searcher actions. Time series analysis is a method often used to understand the underlying characteristics of temporal data in order to make forecasts. In this study, we used a Web search engine transactional log and time series analysis to investigate users’ actions. We conducted our analysis in two phases. In the initial phase, we employed a basic analysis and found that 10% of searchers clicked on sponsored links. However, from 22:00 to 24:00, searchers almost exclusively clicked on the organic links, with almost no clicks on sponsored links. In the second and more extensive phase, we used a one-step prediction time series analysis method along with a transfer function method. The period rarely affects navigational and transactional queries, while rates for transactional queries vary during different periods. Our results show that the average length of a searcher session is approximately 2.9 interactions and that this average is consistent across time periods. Most importantly, our findings shows that searchers who submit the shortest queries (i.e., in number of terms) click on highest ranked results. We discuss implications, including predictive value, and future research.

Groundwater flow model of the Logan river alluvial aquifer system Josephville, South East Queensland

The study focuses on an alluvial plain situated within a large meander of the Logan River at Josephville near Beaudesert which supports a factory that processes gelatine. The plant draws water from on site bores, as well as the Logan River, for its production processes and produces approximately 1.5 ML per day (Douglas Partners, 2004) of waste water containing high levels of dissolved ions. At present a series of treatment ponds are used to aerate the waste water reducing the level of organic matter; the water is then used to irrigate grazing land around the site. Within the study the hydrogeology is investigated, a conceptual groundwater model is produced and a numerical groundwater flow model is developed from this. On the site are several bores that access groundwater, plus a network of monitoring bores. Assessment of drilling logs shows the area is formed from a mixture of poorly sorted Quaternary alluvial sediments with a laterally continuous aquifer comprised of coarse sands and fine gravels that is in contact with the river. This aquifer occurs at a depth of between 11 and 15 metres and is overlain by a heterogeneous mixture of silts, sands and clays. The study investigates the degree of interaction between the river and the groundwater within the fluvially derived sediments for reasons of both environmental monitoring and sustainability of the potential local groundwater resource. A conceptual hydrogeological model of the site proposes two hydrostratigraphic units, a basal aquifer of coarse-grained materials overlain by a thick semi-confining unit of finer materials. From this, a two-layer groundwater flow model and hydraulic conductivity distribution was developed based on bore monitoring and rainfall data using MODFLOW (McDonald and Harbaugh, 1988) and PEST (Doherty, 2004) based on GMS 6.5 software (EMSI, 2008). A second model was also considered with the alluvium represented as a single hydrogeological unit. Both models were calibrated to steady state conditions and sensitivity analyses of the parameters has demonstrated that both models are very stable for changes in the range of ± 10% for all parameters and still reasonably stable for changes up to ± 20% with RMS errors in the model always less that 10%. The preferred two-layer model was found to give the more realistic representation of the site, where water level variations and the numerical modeling showed that the basal layer of coarse sands and fine gravels is hydraulically connected to the river and the upper layer comprising a poorly sorted mixture of silt-rich clays and sands of very low permeability limits infiltration from the surface to the lower layer. The paucity of historical data has limited the numerical modelling to a steady state one based on groundwater levels during a drought period and forecasts for varying hydrological conditions (e.g. short term as well as prolonged dry and wet conditions) cannot reasonably be made from such a model. If future modelling is to be undertaken it is necessary to establish a regular program of groundwater monitoring and maintain a long term database of water levels to enable a transient model to be developed at a later stage. This will require a valid monitoring network to be designed with additional bores required for adequate coverage of the hydrogeological conditions at the Josephville site. Further investigations would also be enhanced by undertaking pump testing to investigate hydrogeological properties in the aquifer.

Land use at privatized Australian airports : classification and analyses

In recent years the air transport industry has experienced unprecedented growth, driven by strong local and global economies. Whether this growth can continue in the face of anticipated oil crises; international economic forecasts and recent influenza outbreaks is yet to be seen. One thing is certain, airport owners and operators will continue to be faced with challenging environments in which to do business. In response, many airports recognize the value in diversifying their revenue streams through a variety of landside property developments within the airport boundary. In Australia it is the type and intended market of this development that is a point of contention between private airport corporations and their surrounding municipalities. The aim of this preliminary research is to identify and categorize on-airport development occurring at the twenty-two privatized Australian airports which are administered under the Airports Act [1996]. This new knowledge will assist airport and municipal planners in understanding the current extent and category of on-airport land use, allowing them to make better decisions when proposing development both within airport master plans and beyond the airport boundary in local town and municipal plans.

Addressing issues in sparseness, ecological bias and formulation of the adjacency matrix in Bayesian spatio-temporal analysis of disease counts

The main objective of this PhD was to further develop Bayesian spatio-temporal models (specifically the Conditional Autoregressive (CAR) class of models), for the analysis of sparse disease outcomes such as birth defects. The motivation for the thesis arose from problems encountered when analyzing a large birth defect registry in New South Wales. The specific components and related research objectives of the thesis were developed from gaps in the literature on current formulations of the CAR model, and health service planning requirements. Data from a large probabilistically-linked database from 1990 to 2004, consisting of fields from two separate registries: the Birth Defect Registry (BDR) and Midwives Data Collection (MDC) were used in the analyses in this thesis. The main objective was split into smaller goals. The first goal was to determine how the specification of the neighbourhood weight matrix will affect the smoothing properties of the CAR model, and this is the focus of chapter 6. Secondly, I hoped to evaluate the usefulness of incorporating a zero-inflated Poisson (ZIP) component as well as a shared-component model in terms of modeling a sparse outcome, and this is carried out in chapter 7. The third goal was to identify optimal sampling and sample size schemes designed to select individual level data for a hybrid ecological spatial model, and this is done in chapter 8. Finally, I wanted to put together the earlier improvements to the CAR model, and along with demographic projections, provide forecasts for birth defects at the SLA level. Chapter 9 describes how this is done. For the first objective, I examined a series of neighbourhood weight matrices, and showed how smoothing the relative risk estimates according to similarity by an important covariate (i.e. maternal age) helped improve the model’s ability to recover the underlying risk, as compared to the traditional adjacency (specifically the Queen) method of applying weights. Next, to address the sparseness and excess zeros commonly encountered in the analysis of rare outcomes such as birth defects, I compared a few models, including an extension of the usual Poisson model to encompass excess zeros in the data. This was achieved via a mixture model, which also encompassed the shared component model to improve on the estimation of sparse counts through borrowing strength across a shared component (e.g. latent risk factor/s) with the referent outcome (caesarean section was used in this example). Using the Deviance Information Criteria (DIC), I showed how the proposed model performed better than the usual models, but only when both outcomes shared a strong spatial correlation. The next objective involved identifying the optimal sampling and sample size strategy for incorporating individual-level data with areal covariates in a hybrid study design. I performed extensive simulation studies, evaluating thirteen different sampling schemes along with variations in sample size. This was done in the context of an ecological regression model that incorporated spatial correlation in the outcomes, as well as accommodating both individual and areal measures of covariates. Using the Average Mean Squared Error (AMSE), I showed how a simple random sample of 20% of the SLAs, followed by selecting all cases in the SLAs chosen, along with an equal number of controls, provided the lowest AMSE. The final objective involved combining the improved spatio-temporal CAR model with population (i.e. women) forecasts, to provide 30-year annual estimates of birth defects at the Statistical Local Area (SLA) level in New South Wales, Australia. The projections were illustrated using sixteen different SLAs, representing the various areal measures of socio-economic status and remoteness. A sensitivity analysis of the assumptions used in the projection was also undertaken. By the end of the thesis, I will show how challenges in the spatial analysis of rare diseases such as birth defects can be addressed, by specifically formulating the neighbourhood weight matrix to smooth according to a key covariate (i.e. maternal age), incorporating a ZIP component to model excess zeros in outcomes and borrowing strength from a referent outcome (i.e. caesarean counts). An efficient strategy to sample individual-level data and sample size considerations for rare disease will also be presented. Finally, projections in birth defect categories at the SLA level will be made.