Parameter estimation in a subcritical percolation model with colouring, or cross-contamination rate estimation for microfluidics
We have just published a research paper1 about a nice combination of statistics and applied probability theory. The story opens with a novel mathematical model to quantify unwanted cross-contamination within lab-on-a-chip microfluidic devices. We apply the method of simulated moments (MSM) to estimate the parameters of the model. This method uses computer simulation with many different parameter proposals until the simulated data is close enough to the observed data. Our main contribution is that we prove the statistical consistency (`correctness’) of this method. Proving that the estimate will converge with probability 1 to the true parameter value as the sample size tends to infinity was a challenge because of the dependence among our random sample’s variables.
My current subject is the seasonality in the number of births. I have looked up the number of live births in several countries over a number of years with monthly resolution. My primary question is whether there are certain months of the year in which more babies are born than could be expected.
There are a few other related, truly intriguing questions, but we will not be able to say much about them based on the available data beyond mere speculation. The most directly accessible one is the seasonality in conception rate: to what extent is it defined by seasons (time spent indoors and the like), and how much influence does culture have on it? What would also be very interesting to find out is the seasonality in the number of sexual intercourses. Well, I can’t say much about the latter one, but I will explain why. Continue reading
I’m somewhat obsessed with the heights of people around me. I also have an interest in harmony and aesthetics. So I have been thinking what height proportions make a nice couple. If you are looking for a partner from the opposite sex (whether for a romantic relationship or for dancing), what height would be a good match to yours?
What follows is going to involve data analysis but there will be nothing inherently scientific about it. Continue reading