Need help with a stats routine

[BobbyG]

Anyone got time to write me a Bayes 2x2 statistics Javascript module?

4 inputs, followed by 6 output cells. Need a "clear/reset" and seed it with overridable default values.

I need it to insert in a public service web page located at http://www.bgladd.com/Total_Information_Awareness

There's a downloadable Excel sheet on the site, but what would be better would be an online JS Bayes 2x2 calculator.

I've looked at some Bayes JS apps on the web, they're not exactly what I'd prefer (too techie).

I don't have time to fiddle with doing an error-proofed form, I've never done Javascript. I do SAS and FoxPro programming at work, so coding the math is not the problem. If I had drag & drop JS editing capability, I'd have already done this.

I'll give you a byline (if you wish).

BG

[khalidali63]

lol...couldn' agree more with Dave,
yes please explain your question more

[BobbyG]

Hi -

See the math on http://www.bgladd.com/Total_Information_Awareness

It's about "Bayesian" statistics, a way to calculate likely future misclassification rates, given a few prior ("anterior") empirical or hypothetical assumptions about false positive rates, false negative rates, and the "prevalence" or proportion of true positives in a population.

A Bayes 2x2 is simply composed of [1] true positive [2] false positive [3] true negative and [4] false negative outcomes, expressible as both/either numbers and/or rates.

All of which are explained on the web page. Basically a refutation of the U.S. Government proposal to try to find terrorists by searching through EVERYONE'S personal private records of every type. The Bayesian question is "how many people are likely to be misclassified?"

How would YOU like it if YOU were mistakenly tagged as a terror suspect after the government went through all your personal records?

I've got a form assembled now, but haven't had time to fiddle with embedding I/O cells and javascript computations in it.

BG