update

_posts/2024-07-01-asap.md

@@ -0,0 +1,23 @@
+---
+title: Datasheet for ASAP
+excerpt: Matlab package for scaling pairwise comparison results (scaling, confidence intervals, statistical testing).
+author:
+  - Aliaksei Mikhailiuk
+  - Rafał Mantiuk
+license: MIT
+tags:
+  - Matlab
+  - Python
+  - open-source
+  - pairwise comparisons
+  - active sampling
+  - adaptive procedures
+categories:
+  - Subjective Test Software
+external_link: https://github.com/gfxdisp/asap/tree/master
+direct_download_link: https://github.com/gfxdisp/asap/archive/refs/heads/master.zip
+---
+
+A toolkit for actively pairing conditions in pairwise comparison preference aggregation.
+
+Pairwise comparison data arise in many domains with subjective assessment experiments. In these experiments participants are asked to express a preference between two conditions. However, many pairwise comparison protocols require a large number of comparisons to infer accurate scores, which may be unfeasible when each comparison is time-consuming or expensive. To address this problem we propose ASAP, an active sampling algorithm, offering the highest accuracy of inferred scores compared to the existing methods. Unlike most existing methods, which rely on partial updates of the posterior distribution, we are able to perform full updates and, therefore, much improve the accuracy of the inferred scores.

_posts/2024-07-01-pwcmp.md

@@ -0,0 +1,26 @@
+---
+title: Datasheet for pwcmp
+excerpt: Matlab package for scaling pairwise comparison results (scaling, confidence intervals, statistical testing).
+author:
+  - Rafał Mantiuk
+  - María Pérez Ortiz
+license: MIT
+tags:
+  - matlab
+  - open-source
+  - pairwise comparisons
+categories:
+  - Quality Analysis
+  - Subjective Test Software
+external_link: https://github.com/mantiuk/pwcmp
+direct_download_link: https://github.com/mantiuk/pwcmp/archive/refs/heads/master.zip
+---
+
+This is a set of Matlab functions for scaling pairwise comparison experiment results based on Thurstone's model V assumptions.
+
+The main features:
+
+* The scaling can work with imbalanced and incomplete data, in which not all pairs are compared, and some pairs are compared more often than others.
+* Additional priors reduce bias due to the non-linear nature of the problem.
+* Outlier rejection to screen observers that perform differently than the rest.
+* The code can compute confidence intervals using bootstrapping.