@jess remove the 0.1*n highest and 0.1*n lowest data points from the sample.
that's usually good enough for this kind of noisy data


@jess I don't think it has a name. The jargon would be "removing outliers by dropping the highest and lowest deciles".
(or "10 highest and 10 lowest percentiles"; adapt if you use other constants than 0.1)

· · Web · 0 · 0 · 0
Sign in to participate in the conversation

oc.todon.fr is a mostly French-speaking Mastodon instance with an active moderation. oc.todon.fr est une instance Mastodon principalement francophone et avec une modération active.