# Neil G

 2d comment What is the Dirichlet equivalent of a Beta (1,1) distribution?Yes, but what if you had used odds $o \in [0, \infty)$ instead of probability $p$. Would you not want to choose a uniform prior for that? 2d accepted Does the principle of indifference apply to the Borel-Kolmogorov paradox? Dec 5 comment How to resolve Simpson's paradox?@Potato: I learned it from his book "Probabilistic Reasoning in Intelligent Systems". There are many tutorials online, but it's hard to find one that builds intuition rather than just presenting the algorithm. Dec 5 comment What is the Dirichlet equivalent of a Beta (1,1) distribution?"this would reflect we know nothing about parameter $p$" — it is a uniform prior, but it's not an uninformative prior. Dec 5 awarded Nice Answer Dec 5 revised How to resolve Simpson's paradox?deleted 48 characters in body Dec 5 revised How to resolve Simpson's paradox?added 26 characters in body Dec 5 revised How to resolve Simpson's paradox?added 1107 characters in body Dec 5 revised How to resolve Simpson's paradox?added 525 characters in body Dec 5 comment How to resolve Simpson's paradox?Good answer, but randomization is presumably impossible. If you can give random treatments to random people, you're done. I added an answer about the front door and back door methods, which as far as I know are the only ways of trying to establish causality based on observational data. Also, "the full table of results" is not necessary. Pearl gives an algorithm that returns the minimum number of variables that need to be measured and used for partitioning. Dec 5 revised How to resolve Simpson's paradox?added 53 characters in body Dec 5 comment How to resolve Simpson's paradox?@John: if you conclude that x->weight is negative because age is correlated with both wegiht and x, doesn't that make it a confounder? Dec 5 answered How to resolve Simpson's paradox? Dec 3 comment Throwing the fattest people off of an overloaded airplane.@user2548100 The computational complexity measures asymptotic runtime. It can be compared between algorithms. Dec 1 comment What is the purpose of the first test in an inductive proof?IH: All sets of $n \ge 4$ lines on the plane intersect at a single point. If it holds true for $n\ge 4$ points, then it holds true for $n+1$ points since the first $n\ge 3$ points intersect at a point and so too do the last $n$ lines and the two intersection points must be the same point. Therefore, all lines on the plane intersect at a single point. Nov 29 awarded Nice Question Nov 25 comment Throwing the fattest people off of an overloaded airplane.@user2548100: If you don't think so, don't vote for my answer. Nov 25 comment Throwing the fattest people off of an overloaded airplane.@user2548100: what are you talking about? I was just illustrating an algorithm. You can code it in C++ if you like. What is the point of this rant or yours? Nov 25 comment Throwing the fattest people off of an overloaded airplane.@user2548100: Like C++, Python makes asymptotic guarantees and so hides as much as say . If anything C++ obscures the algorithm by miring you in architectural details of data types, loop counters, etc. Nov 25 comment Throwing the fattest people off of an overloaded airplane.@user2548100: Do you see anything other than appending elements at the array's end in the above algorithm?