Quantile Regression with longitudinal data

Specialeforsvar ved Emil Hvitfeldt Hansen

Titel: Quantile Regression with longitudinal data

Abstract: It is well known that one of the problems with ordinary OLS regression is the it can be hard to identify how the covariates influences the location, scale, and shape of the entire response distribution. This thesis looks at quantile regression and its way of estimating cthe conditional median, and other quantiles, of the response variable. Which will we enable us to gather much deeper understanding on the effect of the covariances then by ordinary OLS. Furthermore with the introduction of longitudinal data will we talk about a new method to facilitate efficient parameter estimation with the use of Cholesky decomposition. This is appompanied with a simulation study and practical analyses on weather data

 

Vejleder: Helle Sørensen
Censor:   Claus Dethlefsen, Novo Nordisk