Formerly at Lancaster University
My research investigates the application of active model combination methods such as Boosting and Bagging to time series data. It seeks to narrow the gap between the application of combination methods developed in machine learning and the use of these techniques in the forecasting of time series data which is traditionally approached through the use of statistical models. Some expected contributions are:
My research interests include but are not limited to:
Active model combination - an evaluation and extension of Bagging and Boosting for time series forecasting
Since the seminal work by Bates and Granger (1969), the practice of combining two or more models, rather than selecting the single best, has consistently been shown to lead to improvements in accuracy. In forecasting, model combination aims to find an optimal weighting given a set of precalculated forecasts. In contrast, machine learning includes methods which simultaneously optimise individual models and the weights used to combine them. Bagging and boosting combine the results of complementary and diverse models generated by actively perturbing, reweighting and resampling training data. Despite large gains in predictive accuracy in classification, limited research assesses their efficacy on time series data. This thesis provides a critical review of the combination literature, and is the first literature survey of boosting for time series forecasting. The lack of rigorous empirical evidence on forecast accuracy of Bagging and boosting is identified as a major gap. To address this, a rigorous evaluation of Bagging and boosting adhering to recommendations of the forecasting literature is performed using robust error measures on a large set of real time series, exhibiting a representative set of features and dataset properties. Additionally there is a narrow focus on marginal extensions of boosting, and limited evidence of any gains in accuracy. A novel framework is proposed to explore the impact of varying boosting meta-parameters, and to evaluate the empirical accuracy of the resulting 96 boosting variants. The choice of base model and combination size are found to have the largest impact on forecast accuracy. Findings show that boosting overfits to noisy data, however no existing study investigates this crucial issue. New noise robust boosting methods are developed and evaluated for time series forecast models.
PhD in Management Science
Department of Management Science, Lancaster University Management School, Lancaster University, UK
Model selection and combination for time series forecasting with artificial neural networks
Dr. Sven F. Crone, Professor Robert Fildes
MSc (Honors) Computer Science
Department of Computer Science, University of Canterbury, Christchurch, New Zealand
BSc (Honors) Computer Science and Mathematics
Department of Computing and Information Technology, Department of Mathematics and Statistics, University of the West Indies, St. Augustine Campus, Trinidad and Tobago
ACCA Chartered Accountant (Affiliate)
ACCA, London, United Kingdom
www.devonbarrow.com
Academic Work Experience
01/2013 - present Main activities and responsibilities
Main activities and responsibilities
Main activities and responsibilities 01/2007 – 03/2008 Main activities and responsibilities
|
Post Doctoral Researcher Forecasting and statistics postdoctoral research including applied academic research with companies in the field of forecasting, supply chain, marketing analytics and data mining.
Part-time Research Assistant Applied academic research with companies such as: 1) Retail Express Teaching undergraduate courses in database management excel spreadsheet modelling, management science and finance.
Programming and project management support for research projects within the group specializing in artificial intelligence in education, including the design and implementation of intelligent tutoring system. |
Industry Work Experience
03/2008 – 07/2009 Main activities and responsibilities
07/2004 – 01/2006 Main activities and responsibilities
10/1999 – 08/2001
|
Staff accountant II
Staff Accountant 1
Project Administrator Planning, Implementation and Coordination of Project activities including completion of the National Greenhouse Gases Inventory and the National Climate Change Strategy and Adaptation Plan.
|
Research output: Contribution to Journal/Magazine › Journal article › peer-review
Research output: Contribution to Journal/Magazine › Journal article › peer-review
Research output: Contribution to Journal/Magazine › Journal article › peer-review