How Many Do We Need? Exploration of the Population Size Effect on the Performance of Forensic Speaker Classification

Shunichi Ishihara; Yuko Kinoshita

How Many Do We Need? Exploration of the Population Size Effect on the Performance of Forensic Speaker Classification

Research output: Contribution to conference › Paper

Abstract

This paper investigates how change in population size affects the reliability of the likelihood ratio (LR)-based forensic speaker classification. Using features of the long term F0 distribution, we performed LR-based speaker classification and examined its performance with population sizes from 10 to 120. The results revealed that LRs could be heavily influenced by the population data. We discovered that the reliability of LR-based evaluation of the evidence was heavily compromised if the population data was limited to a small number of speakers.

Original language	English
Pages	1941-1944
Publication status	Published - 2008
Event	INTERSPEECH 2008-9th Annual Conference of the International Speech Communication Association - Brisbane Australia Duration: 1 Jan 2008 → …

Conference

Conference	INTERSPEECH 2008-9th Annual Conference of the International Speech Communication Association
Period	1/01/08 → …

Access to Document

http://hdl.handle.net/1885/19127

Cite this

@conference{8fea5458b213429891244876ae6113cb,

title = "How Many Do We Need? Exploration of the Population Size Effect on the Performance of Forensic Speaker Classification",

abstract = "This paper investigates how change in population size affects the reliability of the likelihood ratio (LR)-based forensic speaker classification. Using features of the long term F0 distribution, we performed LR-based speaker classification and examined its performance with population sizes from 10 to 120. The results revealed that LRs could be heavily influenced by the population data. We discovered that the reliability of LR-based evaluation of the evidence was heavily compromised if the population data was limited to a small number of speakers.",

author = "Shunichi Ishihara and Yuko Kinoshita",

year = "2008",

language = "English",

pages = "1941--1944",

note = "INTERSPEECH 2008-9th Annual Conference of the International Speech Communication Association ; Conference date: 01-01-2008",

}

TY - CONF

T1 - How Many Do We Need? Exploration of the Population Size Effect on the Performance of Forensic Speaker Classification

AU - Ishihara, Shunichi

AU - Kinoshita, Yuko

PY - 2008

Y1 - 2008

N2 - This paper investigates how change in population size affects the reliability of the likelihood ratio (LR)-based forensic speaker classification. Using features of the long term F0 distribution, we performed LR-based speaker classification and examined its performance with population sizes from 10 to 120. The results revealed that LRs could be heavily influenced by the population data. We discovered that the reliability of LR-based evaluation of the evidence was heavily compromised if the population data was limited to a small number of speakers.

AB - This paper investigates how change in population size affects the reliability of the likelihood ratio (LR)-based forensic speaker classification. Using features of the long term F0 distribution, we performed LR-based speaker classification and examined its performance with population sizes from 10 to 120. The results revealed that LRs could be heavily influenced by the population data. We discovered that the reliability of LR-based evaluation of the evidence was heavily compromised if the population data was limited to a small number of speakers.

M3 - Paper

SP - 1941

EP - 1944

T2 - INTERSPEECH 2008-9th Annual Conference of the International Speech Communication Association

Y2 - 1 January 2008

ER -

How Many Do We Need? Exploration of the Population Size Effect on the Performance of Forensic Speaker Classification

Abstract

Conference

Access to Document

Fingerprint

Cite this