Hi, I’m Abraham Flaxman, and this blog is about TCS and ACO applications that I’m learning about at the Institute for Health Metrics and Evaluation at UW. I hope to share the inspiring and challenging research directions I’ve found with other CS, Math, and OR people.
The picture in the title bar was adapted from a photo by ninoz, and, like everything else on this blog, has some rights reserved. The favicon is a colored-in version of a public domain image I got from Wikimedia.
Healthy Algorithms 
.jpg){kind=link}
Abraham, I am researching my geneology and I am wondering if you are related to an Abram Flaxman of Russia who was a tailor and settled in Boston in 1907. I am totally obsessed and appreciate your reply. julie713@cox.net
Julie: I don’t think so, my father’s father Abraham lived in New York. He died young, and I don’t know much about him, but I know he was not a tailor. He worked in some sort shipping, maybe more on the stevedore side of things than OR/MS.
Hi Abraham, I wonder if you can help me. I’m an Msc Student @ Leeds University undertaking my project in tagging medical concepts with verbal autopsies. Please see my blog http://mscgirl.wordpress.com/. I was looking for anonymised verbsal autopsy data and found some on your site, which was v.interesting. I took the data and loaded it into a machine learning tool (WEKA) but unfortuantely it didn’t mean to much to me as i could not ascertain what the symptoms and cause of death were, as they were numeric. Is it possible that you could advise me of this information. It would very much help me with my project as with your permission I would like to use your data to explain my technique of concept extraction and classification that I have developed. Thanking you in advance, Rebecca
Hi Rebecca,
I’m working on getting the full data released publicly for people like you to use. But it might take a while…
I’ll move this conversation to the verbal autopsy challenge page, but I wanted to reply here to make sure you got it.
Thanks Abraham for your quick response. My project needs to conclude by end of August of this year. So i guess it will not be fully available by then for me to use. Although you nver know 🙂 In the meatime I will try an puruse some other avenues. Thanks for your post and your article I found it a real interesting read.
Thank you Abraham, I never saw your reply of 2 years ago until now! So no connection to Massachusetts that you know of..I will keep looking – Julie
julie713@cox.net
Abie! Was googling some questions about gaussian processes in python and landed on your blog post about it! Saw a line about “artificial intelligence for health metricians” and thought…I know that class! Hope all is well!
-Ian
Hi Ian!
Dear Abraham
Your blog is a wonderful collection of information on PyMC and MCMC in general, a great help in (re)learning statistics. Thank You!
As e-mail address scrambler on Your page is not working, I’ll ask a question here, if You don’t mind. I believe that MCMC is a fitting tool for this type of problem but I haven’t found a good solution yet.
There is a “black box”/experiment yielding a random sample x[i] when given a set of parameters a[j]. The number of randoms is generally less than a 100, the number of variables is between 4 and 8.
I would like to use MCMC sampler to explore parameter space of a[j] to try and find a set that produces a random sample with distribution as close to a Poisson distribution with a specified rate as possible.
So far I’ve tried PyMC at fitting the sample histogram rather than the distribution itself: produce a desired histogram with numpy.random.poisson(rate,N); set it as an observed data; try to fit it with a histogram produced by the “black box” using parameters a. This “almost works”, but different runs often produce very different results as can be expected for a lowish N. But since there are no actual observed data, only a sample from a distribution with unknown properties, I would like to find another solution. What, in Your opinion, would be a “proper” way to compare these random samples and to minimize their difference? I would be very grateful for even a small nudge in the direction of a better answer. Thank You again!
Respectfully, Dmitrij
Hi Dmitrij. I don’t understand the problem you are trying to solve, or your proposed solution. You should post it to cross-validated, and then lots of smart people will see it (and maybe one will be less busy than me :P).
Here is my go-to advice for posting questions to stack-overflow-type forums: http://stackoverflow.com/help/mcve
Doctor Abraham Flaxman:
With admiration I have started reading your Bayesian meta regression model applied in GBD. At the moment I am conducting a study of burden of disease at hospital level in specific trauma. I respect fully request bibliographic guidance when meta-regression is applied in an individual pathology at the institutional level.
Thank you
Néstor Suarez
Bogotá. Colombia.
Dear Abraham,
I have just implemented the conversion of the data from the WHOVA 2016 questionnaire (implemented in the World Bank’s Survey Solutions CAPI platform) to SmartVA input format.
I found a bug in the SmartVA-Analyze (Smart VA 2.2.0) app:
For the variable adult_5_2 (kind of injury or accident) a value of 7 (Violence, suicide, homicide, abuse) always returns “Undetermined” as the cause of death. All other values (1,2,3,4,5,6,11) return the expected cause.
In Sub-Saharan Africa we currently have a lot of people killed by terrorists, whose cause of death is therefore not detected by SmartVA.
I have some more observations, but would first like to see if this is the appropriate channel to report bugs and make suggestions on SmartVA-Analyze.
Klaus Blass