Monthly Archives: March 2006

Conspiracy Theory?

For those who like F911, Loose Change is definitely a must see documentary giving benefit of doubt about the truth of terrorist attack.

Posted in Entertainment | Leave a comment

Five year anniversary

One day in history reminds my wife and me that it’s been five years since we made our commitment to the marriage at New Haven, CT. We did not exchange wedding rings nor did we have a feast at Yale. … Continue reading

Posted in News and politics | 3 Comments

Congratulations to Dr. Zhao

A bright and deligent Ph.D. student, Zhanlue (George) Zhao defended his dissertation at last. Yes, he is considerate, which means he brought pizza and drink to his thesis committee members for lunch. He is serious about the job, which means he … Continue reading

Posted in News and politics | Leave a comment

10 reasons that I am afraid of being a statistician

10. I consider p-value a deviation from normal.9. I seldom feel complete, sufficient, and consistent.8. I do not like to talk 95% of the time.7. There is a more than 5% chance someone will claim that my work has flaws almost surely.6. I do not know what to do if posterior … Continue reading

Posted in Uncategorized | 2 Comments

It comes at last!

Elevators of the engineering building are now working and we have running water in the entire building! At least I can request to have my circuits lab in one place now. 🙂

Posted in News and politics | Leave a comment

penalty based vs. sample based model selection

发信人: chenhm (老母~我往南方去), 信区: EE标  题: Re: 诚心请教发信站: Unknown Space – 未名空间 (Fri May 20 23:46:45 2005) WWW-POST 【 在 ilovemusic (给情人的摇蓝曲) 的大作中提到: 】: If there are enough data, the following two ways to choose an optimal: regression model:: (1). classify the original data set into a modeling and a validation set.: choose the optimal model with the validation set (such as using PRESS): : (2). use the whole set to develop the model and choose the optimal one with: the information criterion(like AIC or MDL): : which way is better, (1) or (2) ?: Thanks.Just curious: You want to determine the order of linear regression models?What is the purpose of model selection (e.g., for prediction/interpolation)?The problem formulation is unclear to me. —素甘淡泊心常泰  曾履忧危体愈坚 Now looking back, I really do not have a definite answer. I guess for parametric model, penalty based model selection criterion, e.g., MDL, is viable while for nonparametric model, sample based (distribution free) model selection criterion is … Continue reading

Posted in science and engineering | Leave a comment

uncertainties of the upcoming cut

At faculty concil meeting, the Chancellor anticipated that cuts will involve all segments of the operating budget including faculty lines. The non-academic units will be cut. Athletics will be cut. The cuts will be strategic, not random.  The governance method … Continue reading

Posted in News and politics | Leave a comment