I'm a student wanting to write a report in something that has to do with mathematical knowledge. I came on this website on saw an answer by Stefan Hansen, which I would like to implement in my paper.
Counting processes and martingales are objects I view as purely mathematical/probabilistic objects. Nevertheless, they fundamental objects when describing the theory of survival analysis - survival analysis being a branch that is used in many registry-based studies in e.g. epidemiology.
A simple model (a model without censoring) of survival analysis is the following: Let X1,…,Xr be iid random variables with values in (0,∞), where Xi is the lifetime of the ith individual. Let Xi have density f and distribution function F with F(t)<1 for all t∈(0,∞). Put Nit=1{Xi≤t},i=1,…,r, and Nt=∑i=1rNit, i.e. Nt is the number of individuals dead before t. Then (N1t,…,Nrt)t≥0 is an r-dimensional counting process and (Nt)t≥0 is a counting process. Now, theory of local martingales and predictable covariation can be used to derive estimators such as the Nelson-Aalen estimator of the cumulative hazard rate, i.e. the function Λ(t)=−logS(t), where S(t)=1−F(t) is the survival function.
I was wondering if one could explain this at a middle school level. And it would be great if I could get into contact with Mr. Hansen.
After talking with @SimonS, I decided to reinstate my question. I want to show some easy pure mathematics that is not at first sight applied, but can be applied (survival analysis, for example). I want to actually show the maths, but this needs to be readable at the high school level. Thanks again.
Thanks,