Metrics 0.3.0 ndcg

Lodestar.Metrics 0.3.0. This page is frozen at that release. Read the current documentation for what main says now. A link to a decision or a migration page follows main, and leaves the archive.

Ndcg

The one to report. Each row's discounted gain is divided by the gain of its own perfect ranking, so the result is in [0, 1] regardless of how relevance was scaled and rows with different judgement scales can be averaged together.

The ideal is computed without tie averaging, as scikit-learn computes it: ranking a row by its own relevance leaves ties only between equal gains, which no ordering can separate. A row where nothing is relevant has no ideal to divide by and scores 0.

No logBase, unlike Dcg, because ndcg_score has none — the discount cancels in the ratio when both halves share a base, and scikit-learn shares base 2.

Members

Member	What it does
Ndcg.Score	The mean normalized discounted gain over the rows, in [0, 1].