Judge - Participant Benchmark Repository

Published: 24 January 2019| Version 1 | DOI: 10.17632/pyyct6s7nm.1
Niharika Hubli,


The judge-participant benchmark repository is a collection of score matrices simulated using the Lens Model Framework. It consists of 2-dimensional matrices with evaluation scores assigned by each judge to the assigned participant. Each zipped folder (named after the size of the matrices) consists of the following - 1. 2000 sparse matrices (names starting with 'SM') of varying sparsity by 20 levels 2. 100 full matrices (names starting with "FM') from which the sparse matrices have been systematically derived from 3. Correlation values (names starting with 'Results_corr') that contain values of Kendall's Tau, capturing values of correlation between ranking vectors



Competition, Recommendation System, Data Matrix