Abstract

Lasso represents one of the best examples of the success of modern statistical methodology. Not only it witnesses vigorous development over the past two decades but also motivates in depth study of some powerful concepts and ideas such as shrinkage estimates, regularisation, variable and model selection, oracle property, sparse model, degrees of freedom, optimisation algorithms, sparse pattern recovery etc. In this talk, I will try to bring them together under a coherent paradigm. One of the goals is to demystify things so that they become easier to understand and to use. It is particularly suitable for graduate students and researchers unfamiliar with the area but would like​ to know more about it for his/her own research agenda​. ​  

Online Class: https://join.skype.com/QnG2jAdChPaG