I am a Lecturer in Statistics in the Department of Mathematics in Imperial College London, where I am teaching advanced statistical modelling, graphical modelling, and official statistics. My research lies with computational statistics, with a particular focus on streaming data analysis. My theoretical interests include streaming classification, adaptive filtering, stochastic approximation, graphical modelling, model and variable selection and online non-parametric methods.
I graduated with a BAHons in Mathematics from Pembroke College, Cambridge University in 2003, and then obtained an MSc in Learning from Data from the Informatics School of Edinburgh University, as well as an MSc in Logic and Algorithms from the University of Athens. I completed my PhD in the Institute for Mathematical Sciences at Imperial College London, entitled "A statistical framework for streaming data analysis" and then worked as a Research Fellow at the Statistical Laboratory in Cambridge University. I have so far been involved in private consultancy in the security industry, and in e-commerce, and am the Chief Data Scientist of a start-up in streaming data analytics (www.ment.at).
For more details, please download my CV
STATISTICAL MODELLING II (M3S2/M4S2)
In this course, the linear model is generalized in several directions, and the resulting framework is investigated from a theoretical and practical perspective, in an intention to develop core skills in statistical data analysis. The course is in three parts. Part A: The Generalised Linear Model as a unifying statistical framework. Part B: Random and mixed effects models. Part C: Generalized Additive Models. The R statistical package will be used throughout.
GRAPHICAL MODELLING (M5MS09)
Probabilistic graphical models encode the relationships between a set of random variables, in a manner that relies on networks and graph-theoretic intuitions. Primarily, they encode conditional independence assumptions, whereby A is statistically independent of B conditional on the value of C. Just as conditional probability is one of the pillars of modern probability, conditional independence is critical in statistical modelling. It underlies model specification, and allows us to infer, elicit, and understand correlation structures between unobserved variables, given the value s of variables we already know. This course will entail a variety of material, including discrete mathematics (graph theory), statistical modelling, algorithms and computational aspects, as well as applications, involving real data and actual applications. We will also touch upon abstract questions, such as the difference between causality and correlation.
OFFICIAL STATISTICS (M5MS18)
"Statistics are the mirror through which we view society" (David Hand). How can the well-being of a nation be measured? Has the UK been faring better than Europe during the financial crisis? Were the policies of our government succesful? Sound policy making must rely on evidence, and the task of gathering such evidence reliably across a multitude of individuals, social groups, businesses and types of activities, is monumental. This is an exciting time for official statistics: the raw data are increasingly becoming available for public scrutiny, and recent developments are redefining what constitutes "well-being", "progress", and how they can be measured.
et al., 2013, CASOS: a Subspace Method for Anomaly Detection in High Dimensional Astronomical Databases, Statistical Analysis and Data Mining, Vol:6, ISSN:1932-1864, Pages:53-72
Hand DJ, Anagnostopoulos C, 2013, When is the area under the receiver operating characteristic curve an appropriate measure of classifier performance?, Pattern Recognition Letters, Vol:34, Pages:492-495
et al., 2012, Online Linear and Quadratic Discriminant Analysis with adaptive forgetting for streaming classification., Statistical Analysis and Data Mining, Vol:5, Pages:139-166
Anagnostopoulos C, Adams NM, Hand DJ, 2010, Streaming Covariance Selection with Applications to Adaptive Querying in Sensor Networks, Computer Journal, Vol:53, ISSN:0010-4620, Pages:1401-1414
et al., 2009, Temporally adaptive estimation of logistic classifiers on data streams, Advances in Data Analysis and Classification, Vol:3, Pages:243-261