Angela Noufaily 

          Research Associate in Statistics -- The Open University 

   

              I joined the Open University Statistics Department in October 2007 as a PhD
              student, supervised by Prof. Chris Jones. My PhD thesis entitled "Parametric
              quantile regression based on the generalised gamma distribution" develops
              methodology for parametric quantile regression. This is done through an overall 
              package covering all different stages of statistical modelling and using the 
              generalized gamma distribution as a parametric model. The methodology developed
              is applied to medical and environmental data. The three-parameter genaralized 
              gamma distribution encompasses most known life distributions as special cases
              It also offers more flexibility than simpler models as its two shape parameters  
              model skewness and tailweight more accurately. 
              
              In March 2011, I was appointed a Research Associate at the Statistics Department  
              where I am pursuing a three-years project entitled "Statistical outbreak detection 
              methods for large multiple surveillance systems" with Prof Paddy Farrington 
              (Principal Investigator), Prof. Paul Garthwaite (co-Investigator) and Dr Doyo
              Gragn Enki. In the project, we develop methodology for outbreak detection of 
              infectious diseases. This is done in collaboration with Public Health England
              who provided us with data on weekly reports of more than 3000 infections. 
  
              As part of the project, we improved the algorithm used at Public Health England
              for outbreak detection, by reducing the number of false reports while maintaining, 
              and in some cases, improving the power of detection. The improved algorithm is 
              now in use at the Robert Koch Institute in Berlin and is being implemented at Public 
              Health England in their LabBase automated algorithm for outbreak detection of
              infectious diseases. It is also implmented in R via the statistical package
              "surveillance" and the function "FarringtonFlexible". We are currently interested 
              in improving the algorithm further by modeling reporting delays and incorporating
              them in the system. 

             To watch a video on our current work, please click on 
              http://www.youtube.com/watch?v=4i4BWbFp0GY&feature=youtu.be