Need formula for min error threshold between 2 gaussian funcs with different variance

[plornive]

What is the formula. If the two functions have the same variance, it is easy to derive. I just hate deriving this shit.

I need a basic dichotomizer.

Anyone please? Karmaaaaaaaaaaaaaa.

 

[plornive]

Or just some function in Matlab would be great.

 

[Lumberg]

Oh yeah, like you're gonna find that on Elite.

So are you bragging that you're so smart you wonder about things like this or.......

 

[plornive]

Oh yeah, like you're gonna find that on Elite.

So are you bragging that you're so smart you wonder about things like this or.......

I bet NoDaddyNo/OMGWTFBBQ would know. He probably thinks about error thresholds while he's peeing so he knows exactly how close to the edge of the toilet he can go.

I really just want this formula in a hurry for something.

 

[Ffactor]

Use Zippy Dippy Dang Dang!!

 

[plornive]

If the variance of both gaussians is the same, the formula is:

T = (mean1 + mean2)/2 + (sigma^2)/(mean1 - mean2)*ln(weight1/weight2)

If the sigmas are different it is more difficult and I am in no condition to figure it out. I bet there are some people here that know.

 

[plornive]

w1*p1(T) = w2*p2(T)

Where P1 and p2 are gaussian functions. I want to derive T.

 

[Seashell]

Yep, I bet NDN knows as well..

He strikes me as the "fine line between genius and insanity" kinda guy.

 

[plornive]

Yep, I bet NDN knows as well..

He strikes me as the "fine line between genius and insanity" kinda guy.

Yeah. This is something I expect that he might know. Bump for the OMGWTFBBQ.

 

[OMGWTFBBQ]

Between having to spend some time in the bathroom this morning and then some network issues, I'm not going to be on for a bit - I will come back in a bit and look at this.

I'm flattered you thought of me

 

[plornive]

Awesome b0r.

I think I may be solving it, but time will tell...

 

[plornive]

I think I solved it!

It turned out to be easier than I thought... I just kept making mistakes solving the quadratic formula. I think I always just changed a few signs and messed the whole thing up. It was for segmenting an image.

I think it is almost a certainty that you know more than me about pattern recognition, probability, etc. I would love to extract your knowledge.

 

[OMGWTFBBQ]

I just extracted something awful from my bowels in the bathroom upstairs.

Yeah, I don't know if I know much or not - I know the general ideas - but as for the math, I nearly failed calc in college. I had an A+ for the first half and then never went back until the final and just tried to figure out the test and got an F. And F and an A+ worked out to let me at least pass.

As a result, I understand stuff not in the way that I would had I been taught, but just in my own demented reasoning - so I tend to not have the right words for ideas and things.

I have read a lot of books on the neural networks involved in most visual pattern recognition - as for other types, I like Bayesian analysis.

I was supposed to have a meeting next week for a possible job using my brain instead of this shit job that I have now - but I will be on vacation (well, if you can call visiting my sick grandparents a vacation) - so perhaps after that.
It would allow me to do some fun stuff.

 

[rjl296]

i hat math