tag:blogger.com,1999:blog-18206410.comments2013-04-16T13:16:48.244-05:00Haunted by randomnessS. Phil Kimhttps://plus.google.com/103566929111600492508noreply@blogger.comBlogger11125tag:blogger.com,1999:blog-18206410.post-79251362860290659162013-04-16T13:16:48.244-05:002013-04-16T13:16:48.244-05:00Thanks for the comment, Paul. I didn't explain...Thanks for the comment, Paul. I didn't explain enough why who will accept an organ is important. It is because the LYFT(Life Years gained from Transplant) depends on the matching characteristics of donor/patient. We want to find at least similar patients to the real patients who accepted, in terms of characteristics. <br />And, Of course, discard rate is one of the most important statistics here. But, we exclude the discard cases here because discard happens when OPO gives up. Unfortunately, I don't have any idea on how they decides when to give up. <br />You idea of costs for false negative and false positive is great. However, we don't know how much the costs should be. And, out client don't know either. S. Phil Kimhttps://www.blogger.com/profile/10786626079956019453noreply@blogger.comtag:blogger.com,1999:blog-18206410.post-90077374282864465052013-04-14T13:22:34.025-05:002013-04-14T13:22:34.025-05:00My first thought is to ask why, in the context of ...My first thought is to ask why, in the context of your problem, it is important to know who will accept an organ. I suspect it has to do with the probability an organ goes to waste. If you can turn that into expected costs for false positives and false negatives, you can firm an expected cost function for prediction error and look for the fur that minimizes it over the training sample.Paul Rubinhttps://www.blogger.com/profile/05801891157261357482noreply@blogger.comtag:blogger.com,1999:blog-18206410.post-44744071256007825122013-04-10T11:03:05.350-05:002013-04-10T11:03:05.350-05:00One of my friends complained that he cannot post a...One of my friends complained that he cannot post a comment on my blog. <br />If you have a problem to post your comment, please let me know, thru email or twitter.S. Phil Kimhttps://www.blogger.com/profile/10786626079956019453noreply@blogger.comtag:blogger.com,1999:blog-18206410.post-518005847143331012013-04-07T17:52:16.648-05:002013-04-07T17:52:16.648-05:00In my experience, we in OR do distinguish "ri...In my experience, we in OR do distinguish "risk" and "uncertainty". "Risk" is associated with parameters whose value may be unknown, but whose distribution is known. This is the realm of stochastic programming. A common stochastic programming objective is to maximize an expected value subject to constraints. "Uncertainty" is when parameters have values that are unknown and do not come from known distributions. That is the realm of robust optimization. A common robust optimization objective is to get the best outcome in the face of the worst-case realization of the parameter values.<br /><br />Note that almost any deterministic model has analogs that involve risk or uncertainty. Many, many of these models are of real-world interest. So in general, I would say that stochastic and robust models are harder than deterministic ones. But deterministic ones can certainly be hard enough.Matthew Saltzmanhttps://www.blogger.com/profile/03225420623527632062noreply@blogger.comtag:blogger.com,1999:blog-18206410.post-43787635521194908212013-04-07T15:13:04.943-05:002013-04-07T15:13:04.943-05:00I suspect that the people working with fuzzy sets ...I suspect that the people working with fuzzy sets and fuzzy logic (in the mathematical sense, not the congressional sense) would consider "uncertain" and "random" to be quite different.Paul Rubinhttps://www.blogger.com/profile/05801891157261357482noreply@blogger.comtag:blogger.com,1999:blog-18206410.post-22427445363398469372013-04-05T11:48:47.517-05:002013-04-05T11:48:47.517-05:00If including the security of position and stree of...If including the security of position and stree of researching / grant proposals, it is much more (although may not double yet) than just comparing the amount of salary.<br /> <br />Treat your posdoc as the invest of your tenure, it worths of it. Good luck. Marshal Wanghttps://www.blogger.com/profile/12494043311684278647noreply@blogger.comtag:blogger.com,1999:blog-18206410.post-81766890654183772282012-10-25T23:46:22.252-05:002012-10-25T23:46:22.252-05:00Even though I said that the inflation on publicati...Even though I said that the inflation on publication pushes more fresh PhDs into Post Doc, Post Doc positions rely on projects (especially in academic post doc case). Therefore, increasing number of post Doc is not due to the inflation on publication or tough job market, but due to the increment of projects either in the quantity(number) or in the quality(amount of fund). It would be interesting if I can get the statistics of projects given to academic institutes. S. Phil Kimhttps://www.blogger.com/profile/10786626079956019453noreply@blogger.comtag:blogger.com,1999:blog-18206410.post-50149967095674696612012-10-25T23:43:32.341-05:002012-10-25T23:43:32.341-05:00I was pretty sure that it cannot be doubled. But I...I was pretty sure that it cannot be doubled. But I didn't know it is as small as $250 per year. I made my salary doubled (more than that in fact), and I gonna do so again within two years from now, hopefully. :)S. Phil Kimhttps://www.blogger.com/profile/10786626079956019453noreply@blogger.comtag:blogger.com,1999:blog-18206410.post-28316194200036410982012-10-25T16:22:44.113-05:002012-10-25T16:22:44.113-05:00Your comment about profs *not* doubling their sala...Your comment about profs *not* doubling their salary when they get tenure was a masterpiece of understatement. My reward for tenure (and promotion to associate prof.) was a bump of $250 per year.<br /><br />Your first figure is enlightening. When I was finishing my PhD (1979), post-docs in math were probably fairly rare. I heard of them in physics, but I don't recall knowing any math majors back then who got post-docs.Paul Rubinhttps://www.blogger.com/profile/05801891157261357482noreply@blogger.comtag:blogger.com,1999:blog-18206410.post-50072075653469255292012-10-21T07:51:23.192-05:002012-10-21T07:51:23.192-05:00This is a very intuitive way to explain the main s...This is a very intuitive way to explain the main statistics (mean, variance, stdev). It'd be great if you could continue to post similar examples with intuition for other (more complicated) statistics. Jungpilhttps://www.blogger.com/profile/02203046907764830098noreply@blogger.comtag:blogger.com,1999:blog-18206410.post-35151588684681093152012-10-21T00:29:43.541-05:002012-10-21T00:29:43.541-05:00Sorry. It took time to fix the math expression.
I...Sorry. It took time to fix the math expression. <br />It was ok on Safari, but not on Chrome and Firefox. I didn't try IE. <br />However, now it looks fine on Firefox and Chrome also, I guess it is ok on IE also. S. Phil Kimhttps://www.blogger.com/profile/10786626079956019453noreply@blogger.com