77fa8b4ed2
It was a biased estimator, while the alternative one isn't. |
||
---|---|---|
benches | ||
src | ||
tests | ||
.gitignore | ||
.travis.yml | ||
Cargo.toml | ||
README.md |
average
Calculate the average of a sequence and its error iteratively, using constant
memory and avoiding numerical problems. The calculation can be easily parallelized
by using Average::merge
.
Advantages over naive calculation of average and variance
- Avoids loss of precision due to cancellation.
- Only needs a single pass over the samples, at the cost of a division inside the loop.