This paper is a further development of the paper in which the authors proposed an approach to model building for prediction based on the combination of stepwise logistic regression, information criteria, and the best subset selection. The approach inherited some strong features of the three components mentioned above, in particular it helped to avoid the agonizing process of choosing the "Right" critical p-value in stepwise regression. At the same time automatic selection procedures are often criticized severely for instability and bias in regression coefficients estimates, their standard errors and confidence intervals. To avoid these disadvantages this paper proposes to use some validation, shrinkage and averaging techniques.
Did you find this white paper useful?
11 out of 27 users found this white paper useful
Related white papers
Connecting international businesses securely
Globalisation, efficiency and responsiveness in an incresaingly regulatory environment
Harnessing technology for competitive advantage
The Leisure, Entertainment and Travel Services ICT transition.
PCI Compliance
The new reality for European retail companies
Stakeholders' management and collaboration
How to devise an effective communication strategy
The need for real-time communications in European logistics
Value-added services beyond transportation
e-booking and e-ticketing
How to achieve sustainable competitive advantage in the Leisure, Entertainment and Traval industry.
Live Webcast: Telecoms 2.0 - Where is telecoms heading?
Telecoms 2.0 - Where is telecoms heading? UK telecoms is at a crossroads. IT managers face new demands to enable flexible working, deliver converged networks and provide support for multiple applications...
