# Ram Rachum

less info
reputation
25
bio website ram.rachum.com location Tel Aviv, Israel age member for 3 years, 2 months seen Oct 9 '11 at 23:12 profile views 35

Israeli Python hacker.

I'm the creator of PythonTurtle and the Python Toolbox.

# 22 Actions

 Mar15 awarded Notable Question Dec30 awarded Popular Question Oct8 awarded Yearling Aug12 awarded Supporter Aug12 awarded Scholar Aug12 awarded Nice Question Aug12 comment Counterintuitive result for random sample I assumed a uniform prior. Aug12 comment Counterintuitive result for random sample @mbq: Sure, thanks. (I changed OpenID to ram.rachum.com.) Aug12 awarded Student Aug12 asked Counterintuitive result for random sample Oct18 comment Bayes' Theorem and Agresti-Coull: Will it blend? Wow @whuber, you blew my mind. I have no idea why I've been doing this. I got the impression this is what people do on Bayesian spam filters, so that's what I've been doing. ($P(A|B)$ is the probability a message is spam given that it contains a certain word, for example 'cheese'.) Now I tried to simply calculate $P(A|B)$ directly, and it seems to work, so I don't even need the answer to this question. Unless I'm missing something. Maybe you have an idea why people use Bayes' Theorem in Bayesian spam filtering when filtering by a single word like this? Oct18 accepted What's the accuracy of data obtained through a random sample? Oct18 comment Bayes' Theorem and Agresti-Coull: Will it blend? @Srikant: Agresti-Coull lets you draw conclusion about the true proportion from a small finite sample, and gives you a confidence interval with an answer. I have a similar situation: I have a small finite sample and I want to get the value of $P(A|B)$ along with a confidence interval. To give more context: I'm developing something like a Bayesian spam filter, and this is the probability that a message is spam given that it has a certain word. Oct18 comment Bayes' Theorem and Agresti-Coull: Will it blend? @whuber: For $P(A)$ I check how many of the messages in my finite sample have $A$, and divide that by the size of the sample. For $P(B|A)$ I check how many of the messages that have $A$ also have $B$, and divide that by the number of messages that have $A$. etc. Oct18 asked Bayes' Theorem and Agresti-Coull: Will it blend? Oct11 asked Why break down the denominator in Bayes' Theorem? Oct9 accepted Calculating percentile of normal distribution Oct9 comment Calculating percentile of normal distribution (I have version 7.) I have no problem loading the Statistics package. But what's the function in there called? Because I get the impression that this Quantile line will do the calculation manually instead of using a formula. Oct9 asked Calculating percentile of normal distribution Oct8 comment What's the accuracy of data obtained through a random sample? Again: I am interested in what I can know, not in what I can't know. Let's take your example where $j=0$ and $n=10$. The most likely proportion of redheads is 0%, but there's good chance it's 2% or 5% or 10%. So my question is: Given that $j=0$ and $n=10$, what is the probability distribution function of the proportion of redheads, from the information that I know, not the information that I don't know?