Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!pipex!bnr.co.uk!bnrgate!nrcnet0!cunews!torn!cs.utexas.edu!sdd.hp.com!ux1.cso.uiuc.edu!usenet.ucs.indiana.edu!ectl-sub@snowhite.cis.uoguelph.ca
From: ectl-sub@snowhite.cis.uoguelph.ca (Electronic Communal Temporal Lobe  -  ECTL)
Subject: ECTL 2.10.04
Message-ID: <9210302209.AA24545@snowhite.cis.uoguelph.ca>
Sender: <ectl-sub@snowhite.cis.uoguelph.ca>
Organization: Indiana University
Date: Fri, 30 Oct 1992 22:09:40 GMT
Lines: 123


		Electronic Communal Temporal Lobe
		=================================



Issue: 2.10.04

========================================================================

(1) John Ohala, University of Alberta, Canada; John_Ohala@mts.ucs.ualberta.ca
       Subject: News From The ICSLP 92, Banff

(2) Terrance Nearey, Univ. of Alberta, Canada; nearey@nova.ling.ualberta.ca
       Subject: Re: 2.10.01  Error Rate Estimation

========================================================================

1)

NEWS FROM THE ICSLP 92, BANFF
 
At the International Conference on Spoken Language
Processing, Banff, Alberta, Canada, 12-16 Oct 92 (ICSLP 92)
a PERMANENT COUNCIL FOR THE ORGANIZATION OF ICSLPs
(PCICSLP) was formed, modelled in part on the Permanent
Council for the Organization of Phonetic Sciences
Congresses.  Its sole function is to select the venue of
future ICSLPs and to provide continuity for conferences in
this series.  The actions of this Council are governed by
a constitution.
 
The PCICSLP's initial membership consists of:
 
     Jens Blauert                  Joseph Mariani
     Renato De Mori                John Ohala (President)
     Frank Fallside                Giancarlo Pirani
     Gunnar Fant                   Louis Pols
     Adrian Fourcin                Larry Rabiner
     Hiroya Fujisaki               Mario Rossi
     Wolfgang Hess                 Kenneth Stevens
     John Laver                    Victor Zue
 
The Council accepted a bid from Professor Katsuhiko Shirai
(Waseda University) for the hosting of ICSLP 94 in
Yokohama, Japan.
 
The Council received two bids for the hosting of ICLSP 96
and invites additional such bids to be received no later
than 31 Dec 92.  Bids should specify the availability of a
suitable conference facility (e.g., hotel or convention
center), the personnel who would constitute the organizing
committee, and financial arrangements.
 
The Council invites nominations for additional members of
the Council.  Such nominations should be received no later
than 31 Dec 92.  Election of new members from the slate of
those nominate will be done by the Council in early 1993.
 
Address for correspondence to the PCICSLP:
 
     John Ohala / PCICSLP
     Dept of Linguistics
     Univ of California
     Berkeley, CA 94720 USA
 
    FAX:  +1.510.643 5688
        e-mail:  ohala@cogsci.berkeley.edu

----------------------------------------------------------------------

2)

A couple of references of interest to the problem of error rate estimation:

1) Efron, Bradley. 1982. The jackknife, the bootstrap and other resampling 
plans. Society for Industrial and Applied Mathematics, Philadelphia.
(CBMS-NSF Regional Conference Series in Applied Mathemetics).

2) Hand, D.J. 1981. Discrimination and Classification.
Wiley, New York. See especially p 186-187.

3) Lachenbruch, P.A.  and Mickey, M.R. 1968. Estimation of error rates in 
discriminant analysis. Technometrics 8, 657-662. 

I've forgotten exactly what the original question was, so the rest of this 
note may not be fully relevant. The "leaving one-point out" (as Hand calls 
them, Lachenbruch and Mickey call this the U-method, in the case of linear 
discriminant analysis) method of error rate classification, whereby each 
sample point is classified by a classifier trained on the rest of the data 
is not exactly equivalent to the jackknife, though it bears a resemblence. 
My recollection is that Efron's monograph stresses that trustworthy 
non-parametric point estimators (e.g. the error rate) are easier to come by 
than unbiased confidence intervals using resampling methods. 

Terrance Nearey
U. of Alberta
**************************
Terrance M. Nearey
Department of Linguistics
University of Alberta
Edmonton AB, Canada T6G 2E7
Phone  403 492-5240  Fax 403 0806
email Nearey@nova.ling.ualberta.ca
***************************

=========================
End.


+-----------------------------------------------------------------------+
|									|
|		     Electronic Communal Temporal Lobe			|
|									|
|  Moderator:  David Leip, University of Guelph, Computing & Info. Sci. |
|		Guelph, Ontario, Canada  N1G 2W1.   			|
|		(519) 824-4120  ext.3709 (office) ext.4297 (lab).	|
|									|
|  Send submissions for ECTL to: ectl-sub@snowhite.cis.uoguelph.ca	|
|  Send all other requests to: ectl-request@snowhite.cis.uoguelph.ca	|
|  Anonymous ftp archive at: snowhite.cis.uoguelph.ca  in: pub/ectl     |
|									|
+-----------------------------------------------------------------------+
