"A Two-scale Framework for Variable Selection with Ultrahigh-dimensionality" (Next Week at the Statistics Seminar)
Attention conservation notice: Only of interest if you (1)
have a vast number of variables you could use in your statistical models and
want to reliably learn which ones matter, and (2) are in Pittsburgh in
As always, the seminar is free and open to the public:
- Jianqing Fan, "A Two-scale
Framework for Variable Selection with Ultrahigh-dimensionality"
- Abstract: Ultrahigh-dimensionality characterizes many contemporary statistical
problems from genomics and genetics to finance and economics. We
outline a unified framework to ultrahigh dimensional variable
selection problems: Iterative applications of vast-scale screening
followed by moderate-scale variable selection. The framework is
widely applicable to many statistical contexts: from multiple
regression, generalized linear models, survival analysis to machine
learning and compress sensing.
- The fundamental building blocks are marginal variable screening and
penalized likelihood methods. How high dimensionality can such
methods handle? How large can false positive and negative be with
marginal screening methods? What is the role of penalty functions?
This talk will provide some fundamental insights into these problems.
The focus will be on the sure screening property, false selection size,
the model selection consistency and oracle properties. The advantages
of using folded-concave over convex penalty will be clearly
demonstrated. The methods will be convincingly illustrated by
carefully designed simulation studies and the empirical studies on
disease classifications using microarray data and forecast home price
indexes at zip level.
- Place and time: 4--5 pm on Monday, 12 April, in Porter Hall 125C, CMU
Let add that Fan and
on time series is one of the best available.
Enigmas of Chance
Posted at April 08, 2010 13:45 | permanent link