|
|
@@ -283,7 +283,7 @@ reader, but a probability density function can even be defined in terms of the d
|
|
|
function of `I`, which of course we do not know the distribution of but should tend to
|
|
|
be normal for larger and larger `I`. If that is the case, we can approximate the expectation
|
|
|
of this cosine distribution, where `μ` and `σ` are the mean and variance of the _angle_
|
|
|
-distribution, with [a horrible integral](https://www.wolframalpha.com/input?i=%5Cint_%7B-%5Cinfty%7D%5E%7B%5Cinfty%7D+%5Cfrac%7Bye%5E%7B-%5Cfrac%7B%28%5Cfrac%7By-%5Cmu%7D%7B%5Csigma%7D%29%5E2%7D%7B2%7D%7D%7D%7B%5Csigma%5Csqrt%7B2%5Cpi%7D%5Csqrt%7B1-y%5E2%7D%7D+dy).
|
|
|
+distribution, with [a horrible integral](https://www.wolframalpha.com/input?i=%5Cint_%7B-%5Cinfty%7D%5E%7B%5Cinfty%7D+%5Cfrac%7Bye%5E%7B-%5Cfrac%7B%28%5Cfrac%7Bacos%28y%29-%5Cmu%7D%7B%5Csigma%7D%29%5E2%7D%7B2%7D%7D%7D%7B%5Csigma%5Csqrt%7B2%5Cpi%7D%5Csqrt%7B1-y%5E2%7D%7D+dy).
|
|
|
|
|
|
If this integral was solvable, we could then compare it to the mean of the angle
|
|
|
distribution and see how the error trends.
|
