recommenders-team/recommenders

[ASK] Question on ndcg_at_k calculation

Open

#1,962 opened on Jul 24, 2023

View on GitHub
 (0 comments) (0 reactions) (0 assignees)Python (2,972 forks)batch import
help wanted

Repository metrics

Stars
 (17,706 stars)
PR merge metrics
 (Avg merge 6d 16h) (10 merged PRs in 30d)

Description

Description

Why are we using rank ('first) to get the order of the ideal ranking instead of rank('min') or rank('average)?

https://github.com/microsoft/recommenders/blob/main/recommenders/evaluation/python_evaluation.py#L687 line 597 df_idcg["irank"] = df_idcg.groupby(col_user, as_index=False, sort=False)[ col_rating ].rank("first", ascending=False)

In this case, if there is a tied in the rating, for example, item A, B, C, D, rating is 1, 0, 0,0. Using rank('first), irank = 1,2,3,4. But should we take the tied condition into consideration, that means the irank = 1,2,2,2?

Other Comments

Contributor guide

Research direction
Investigate the impact of using different rank methods (first, min, average) for tied ratings on NDCG calculation. Compare the current implementation with the ideal handling of ties.
Tech stack
python
Domain
databackend
Issue type
Research
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
Half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Fresh
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
PythonRecommendation Systems
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
50