I’m so excited that I’ll see these performers tomorrow. After the concert… by waiting in the line about 30 minutes, I got to get signatures and asked a few questions. Ian was very English and Fabio was very Italian, based on their replys to my questions. I briefly wonder how I would sound when a unexpected asks me questions.

Lecture-wise, it was o.k. Some lecturers were good, some lecturers looked a bit careless. Overall, the lecture quality was somewhat comparable to those summer schools I attended in the US. The rest, it was really bad. This summer school was the worst compared to those summer schools I attended or I helped. Always, small things matter. Once I have time, I’ll list what was good and what were bad. You can expect a long list of bad part but an extremely short list of good part.

1. The reception at the residence hall was too careless. I had to stay longer because transportation schedule didn’t fit to the conference schedule. The housing included the breakfast but they didn’t guide me a place to eat. Later, I asked and they only told a place that is closed.
2. The conference housing never replaced towels during the week stay. It cost 40 pounds and the room was small. No amenities and no fresh towels. DRATS!!!

Women are rejecting marriage in Asia. The social implications are serious”

I’m definately partaking this trend. I’m just scared to take over all the duties that Asian women are expected/supposed to do without outspeaking traditionally while working full time.

Reading it reminds me a popular quote from the art of war:

if you know your enemies and know yourself, you can win a hundred battles without a single loss.

Frankly, I wanted to go pattern recognition summer school but the bureaucratics (to be exact, lack of understanding of needs for methodologies from machine learning for an interdisciplinary research and a bit of igorance and aloofness) hinders me from going.

Anyway, the title is

Accounting for Calibration Uncertainties in X-ray Analysis: Effective Areas in Spectral Fitting

Hyunsook Lee and many others

The full paper is currently available at ArXiv:1102.4610.

Hyunsook Lee, G. Jogesh Babu, and C.R. Rao is accepted.

Not knowing the policy of the publisher, I’m not posting the pdf copy. Let me know if you are interested.

Procedures such as Akaike information criterion (AIC), Bayesian information criterion (BIC), Minimum description length (MDL), and Bootstrap information criterion have been developed in the statistical literature for model selection. Most of these methods use estimation of bias. This bias, which is inevitable in model selection problems, arises from estimating the distance between an unknown true model and an estimated model. Instead of bias estimation, a bias reduction based on jackknife type procedure is developed in this paper. The jackknife method selects a model of minimum Kullback-Leibler distance through bias reduction. It is shown that
(a) the jackknife maximum likelihood estimator is consistent,
(b) the jackknife estimate of the log likelihood is asymptotically unbiased, and
(c) the stochastic order of the jackknife log likelihood estimate is $O(\log \log n).$ Because of these properties,
the jackknife information criterion is applicable to problems of choosing a model from separated families especially when the true model is unknown. Compared to popular information criteria which are only applicable to nested models such as regression and time series settings, the jackknife information criterion is more robust in terms of filtering various types of candidate models in choosing the best approximating model.

Frequentist methods are like a black box.

A astronomer told me so after reading my post “bipartisanship”. Hmm… Another discrepancy. The notion of a black box or its modeling seems different between astronomers and statisticians. Because of my unique education, I do understand both parties. Unfortunately, this bipartisanship makes me hard to go against astronomers lack of statistics education.

Since I couldn’t find astronomer’s account for a black box in analyzing data, I only recommend two papers written by statisticians that explains statisticians’ idea of a black box and the cultural mixture of Bayesian and frequentist statistics, which is not like the mixture of oil and water. Both are solvents of many types of data analysis. Bipartisanship exists although one can claim its superiority to the other under a particular circumstance.

This post was written almost 3 years ago and I don’t remember what papers I specifically pointed at that time. They are most likely by Breiman and by Efron. Particularly the latter scholar clings more. Please, check articles by Efron published in Statistical Science.

We meet people. They come and go. It is very likely that we will never encounter them as our life changes its stage. Atsu graduated and went back to Japan, which country I have no connection to.

Facebook connected us like many others all over the world. Because of its large and broad network structure, the weight of our linkage is mediocre. Yet, it made me find his appearance at a Korean TV documentary aired on the largest broadcasting company (I thinkg KBS is still public).

By the way, the name of the show is “KBS Special: 대지진 100일 – 흔들리는 일본.”

Lastly, it is strange to see him on TV. What’s the odds that a japanese person appears on a Korean National TV program???