#4504 closed Bugs (fixed)
Error in computation of Student's t distribution quantiles
| Reported by: | Owned by: | John Maddock | |
|---|---|---|---|
| Milestone: | Boost 1.44.0 | Component: | math |
| Version: | Boost 1.44.0 | Severity: | Problem |
| Keywords: | Cc: |
Description
When using a students_t_distribution<double> with a non-integer parameter slightly larger than 2, e.g. students_t_distribution<double> s(2.001), then the computation of the quantile(s, p) yields completely wrong results for certain values of p. For example: quantile(s,0.57) yields 536.146. The correct value is somewhere around 0.2. If the degrees of freedom parameter is set to exactly 2 or slightly smaller, the results are correct.
Change History (4)
comment:1 by , 12 years ago
comment:2 by , 12 years ago
Patch above fixes the issue I believe - the issue effects the students_t when
2 < df < 3
I'm not happy that our tests didn't catch this, so I'm leaving this open till that's fully fixed.
comment:3 by , 12 years ago
| Resolution: | → fixed |
|---|---|
| Status: | new → closed |
Note:
See TracTickets
for help on using tickets.
(In [64577]) Fix typo in polynomial approximations. Refs #4504.