weighted_rank_correlation

An implementation weighted rank correlation, that allows to be more flexibly tuned.

https://github.com/kiuml/weighted_rank_correlation

Science Score: 57.0%

This score indicates how likely this project is to be science-related based on various indicators:

✓
CITATION.cff file
Found CITATION.cff file
✓
codemeta.json file
Found codemeta.json file
✓
.zenodo.json file
Found .zenodo.json file
✓
DOI references
Found 2 DOI reference(s) in README
○
Academic publication links
○
Academic email domains
○
Institutional organization owner
○
JOSS paper metadata
○
Scientific vocabulary similarity
Low similarity (6.7%) to scientific vocabulary

Keywords

distance ranking

Last synced: 9 months ago · JSON representation ·

Repository

An implementation weighted rank correlation, that allows to be more flexibly tuned.

Basic Info

Host: GitHub
Owner: KIuML
Language: Python
Default Branch: main
Homepage:
Size: 29.3 KB

Statistics

Stars: 2
Watchers: 2
Forks: 1
Open Issues: 2
Releases: 0

Topics

distance ranking

Created almost 3 years ago · Last pushed over 2 years ago

Metadata Files

Readme Citation

Weighted Rank Correlation

An implementation of Henzgens and Hüllermeiers weighted rank correlation coefficient scaled gamma, provided by Hoang Cong Thanh @Marmalada1.

Installation

The project can be installed with poetry

bash poetry add git+https://github.com/KIuML/weighted_rank_correlation.git

Usage

To perform a calculation run the function gamma_corr(x,y) with x and y being the two rankings to compare.

Consider the following minimal example for usage:

```python from gammacorrelation import genweights, gamma_corr

rankinga = [1, 2, 3, 4, 5] rankingb = [5, 4, 3, 2, 1]

print(gammacorr(rankinga, ranking_b, weights="top")) ```

This example computes the gamma correlation coefficient for the predifined top weighting function, that emphasizes changes in the upper part of the rankings.

Custom weighting functions can be defined in terms of a vector of length n - 1. The weight at position i describes the degree to which rank position i and i + 1 are distinguished from each other. For example, a linearly inreasing weighting (i.e. emphasizing changes at the bottom of the rankings) can be implemented as follows:

```python from gammacorrelation import genweights, gamma_corr

rankinga = [1, 2, 3, 4, 5] rankingb = [3, 2, 5, 4, 1] w = np.linspace(1, .25, 4)

print(gammacorr(rankinga, ranking_b, weights=w)) ```

Owner

Name: KIuML
Login: KIuML
Kind: organization

Repositories: 3
Profile: https://github.com/KIuML

Citation (CITATION.cff)

cff-version: 1.2.0
message: "If you use this software, please cite it as below."
authors:
- family-names: "Henzgen"
  given-names: "Sascha"
- family-names: "Hüllermeier"
  given-names: "Eyke"
  orcid: "https://orcid.org/0000-0002-9944-4108"
title: "Weighting by Tying: A New Approach to Weighted Rank Correlation"
url: "https://arxiv.org/abs/2308.10622"

ecosyste.ms

Data

Tools

Indexes

Applications

Experiments

Open Source Science