Revisiting Inter-Genre Similarity

Bob L. Sturm, Fabien Gouyon

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

1 Citation (Scopus)
334 Downloads (Pure)

Resumé

We revisit the idea of ``inter-genre similarity'' (IGS)
for machine learning in general,
and music genre recognition in particular.
We show analytically that the probability of error for IGS
is higher than naive Bayes classification with zero-one loss (NB).
We show empirically that IGS does not perform well,
even for data that satisfies all its assumptions.
OriginalsprogEngelsk
TidsskriftI E E E Signal Processing Letters
Vol/bind20
Udgave nummer11
Sider (fra-til)1050-1053
Antal sider4
ISSN1070-9908
DOI
StatusUdgivet - nov. 2013

Fingerprint

Learning systems
Naive Bayes
Music
Machine Learning
Zero
Similarity

Citer dette

Sturm, Bob L. ; Gouyon, Fabien. / Revisiting Inter-Genre Similarity. I: I E E E Signal Processing Letters. 2013 ; Bind 20, Nr. 11. s. 1050-1053.
@article{e2857c73ccd146e0bb93dd8a0ad0695d,
title = "Revisiting Inter-Genre Similarity",
abstract = "We revisit the idea of ``inter-genre similarity'' (IGS)for machine learning in general,and music genre recognition in particular.We show analytically that the probability of error for IGSis higher than naive Bayes classification with zero-one loss (NB).We show empirically that IGS does not perform well,even for data that satisfies all its assumptions.",
author = "Sturm, {Bob L.} and Fabien Gouyon",
year = "2013",
month = "11",
doi = "10.1109/LSP.2013.2280031",
language = "English",
volume = "20",
pages = "1050--1053",
journal = "I E E E Signal Processing Letters",
issn = "1070-9908",
publisher = "IEEE",
number = "11",

}

Revisiting Inter-Genre Similarity. / Sturm, Bob L.; Gouyon, Fabien.

I: I E E E Signal Processing Letters, Bind 20, Nr. 11, 11.2013, s. 1050-1053.

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

TY - JOUR

T1 - Revisiting Inter-Genre Similarity

AU - Sturm, Bob L.

AU - Gouyon, Fabien

PY - 2013/11

Y1 - 2013/11

N2 - We revisit the idea of ``inter-genre similarity'' (IGS)for machine learning in general,and music genre recognition in particular.We show analytically that the probability of error for IGSis higher than naive Bayes classification with zero-one loss (NB).We show empirically that IGS does not perform well,even for data that satisfies all its assumptions.

AB - We revisit the idea of ``inter-genre similarity'' (IGS)for machine learning in general,and music genre recognition in particular.We show analytically that the probability of error for IGSis higher than naive Bayes classification with zero-one loss (NB).We show empirically that IGS does not perform well,even for data that satisfies all its assumptions.

U2 - 10.1109/LSP.2013.2280031

DO - 10.1109/LSP.2013.2280031

M3 - Journal article

VL - 20

SP - 1050

EP - 1053

JO - I E E E Signal Processing Letters

JF - I E E E Signal Processing Letters

SN - 1070-9908

IS - 11

ER -