Timeline for Reducing the error of Algorithms by assigning variables formulas instead of values
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 19, 2013 at 8:10 | vote | accept | Ruhollah Majdoddin | ||
| Apr 19, 2013 at 8:06 | comment | added | Ruhollah Majdoddin | to Emil: ... otherwise A has a bounded error. | |
| Apr 19, 2013 at 7:59 | vote | accept | Ruhollah Majdoddin | ||
| Apr 19, 2013 at 8:10 | |||||
| Apr 19, 2013 at 7:58 | vote | accept | Ruhollah Majdoddin | ||
| Apr 19, 2013 at 7:59 | |||||
| Mar 17, 2013 at 18:45 | answer | added | AlainD | timeline score: 1 | |
| Mar 12, 2013 at 12:15 | comment | added | J.J. Green | Federico Poloni : am I allowed to use log1p() ? :-) | |
| Mar 12, 2013 at 11:56 | comment | added | Emil Jeřábek | Since the algorithm can use conditionals, even a slightest error in one variable may lead to choosing the wrong branch of a conditional, and therefore to a completely different result, regardless of whether you use A or B. | |
| Mar 12, 2013 at 10:51 | comment | added | Federico Poloni | "Less operations" does not always mean "lower error": for instance, compute in a double-precision environment $\frac{\log(1+x)}{x}$ and $\frac{\log(1+x)}{(1+x)-1}$ for $x=10^{-15}$. | |
| Mar 12, 2013 at 10:46 | history | asked | Ruhollah Majdoddin | CC BY-SA 3.0 |