The Annals of Applied Statistics
- Ann. Appl. Stat.
- Volume 13, Number 1 (2019), 492-519.
A Bayesian Mallows approach to nontransitive pair comparison data: How human are sounds?
We are interested in learning how listeners perceive sounds as having human origins. An experiment was performed with a series of electronically synthesized sounds, and listeners were asked to compare them in pairs. We propose a Bayesian probabilistic method to learn individual preferences from nontransitive pairwise comparison data, as happens when one (or more) individual preferences in the data contradicts what is implied by the others. We build a Bayesian Mallows model in order to handle nontransitive data, with a latent layer of uncertainty which captures the generation of preference misreporting. We then develop a mixture extension of the Mallows model, able to learn individual preferences in a heterogeneous population. The results of our analysis of the musicology experiment are of interest to electroacoustic composers and sound designers, and to the audio industry in general, whose aim is to understand how computer generated sounds can be produced in order to sound more human.
Ann. Appl. Stat., Volume 13, Number 1 (2019), 492-519.
Received: December 2017
Revised: June 2018
First available in Project Euclid: 10 April 2019
Permanent link to this document
Digital Object Identifier
Mathematical Reviews number (MathSciNet)
Zentralblatt MATH identifier
Crispino, Marta; Arjas, Elja; Vitelli, Valeria; Barrett, Natasha; Frigessi, Arnoldo. A Bayesian Mallows approach to nontransitive pair comparison data: How human are sounds?. Ann. Appl. Stat. 13 (2019), no. 1, 492--519. doi:10.1214/18-AOAS1203. https://projecteuclid.org/euclid.aoas/1554861658
- Supplement to “A Bayesian Mallows approach to nontransitive pair comparison data: How human are sounds?”. In supplement A the adaptations of the MCMC algorithm to the logistic and finite mixture model extensions (of Sections 3.2 and 3.3) are explained. Supplement B describes the procedure to randomly sample from the proposed model. The procedure was used to generate simulated and nested datasets for Section 6. Supplement C presents results obtained from experiments on simulated data generated from the logistic model for mistakes. Finally, in supplement D, we report diagnostic plots to study convergence and mixing of the MCMC procedure proposed in the paper.