3 ms·
Aren't intervals a rather limited substitute for distributions (of error)? Using the intervals as if they and their joint distributions are uniform doesn’t make
by tgv 2y ago
Aren't intervals a rather limited substitute for distributions (of error)? Using the intervals as if they and their joint distributions are uniform doesn’t make much sense to me. Perhaps if you really need to know the limits, but then as was mentioned elsewhere, the interval widths grow very quickly.
- klysm 2y agoYes they are very limited. The sum of two uniform distributions is not uniform. If you care about the probability density within the interval you need to reach for more advanced methods.