Print

Print


 
Papers of interest are

Efron B -1975- the efficiency of logistic regression compared to normal discriminant analysis , JASA v 70 n 552, pg 892
Efron,B -1987- The effifiency of multinomial logistic regression compared with multiple group dicriminant analysis,
JASA , v 82, n )) pg 1118

Basilio

2012/3/17 Nigel Marriott <[log in to unmask]>

Hi Allstatters,

 

Thank you for the responses to my query (8 in all).  The point that I took out was that DA relies on the assumption of Multivariate Normality of the predictor variables whilst MLR doesn’t make this assumption.  I sensed a mild preference for MLR from respondents with one correspondent noting that MLR was developed after the publication by Mardia, Kent, Bibby on multivariate analysis.  Two referred me to this link http://www-stat.stanford.edu/~tibs/ElemStatLearn/ for more details.  A few were surprised that I didn’t have DA in my version of R though I should have said I am using RExcel so it may not appear in the RExcel menus but is probably still in my R package.

 

However, nobody answered my point (which wasn’t well phrased) as to why Wikipedia fails to make mutual references to these 2 techniques.  If you read these wiki pages http://en.wikipedia.org/wiki/Linear_discriminant_analysis for DA and http://en.wikipedia.org/wiki/Multinomial_logit for MLR, you will note both make copious references to related methods but neither refer to each other which I found odd given the similarity of purpose.  That led me to wonder if I had misunderstood the purpose of these 2 techniques.

 

Regards

 

Nigel Marriott

Chartered Statistician

www.marriott-stats.com

 

Ground Floor, 21 Marlborough Buildings, Bath BA1 2LY, United Kingdom

Tel (mobile)  +44 (0)773 4069997

Tel (office)    +44 (0)1225 489033

Fax                  +44 (0)870 6221969

 

Marriott Statistical Consulting Limited, Company No. 5577275, VAT No. 883304029

Registered in England,   Registered Office – Equity House, 4-6 School Road, Tilehurst, Reading, RG31 5AL

 

You may leave the list at any time by sending the command

SIGNOFF allstat

to [log in to unmask], leaving the subject line blank.




--

Basilio de Bragança Pereira ,DIC and PhD(Imperial College), DL(COPPE)
*UFRJ-Federal University of Rio de Janeiro
*Titular Professor of  Bioestatistics and of Applied Statistics
*FM-School of Medicine and COPPE-Posgraduate School of Engineering and
HUCFF-University Hospital Clementino Fraga Filho.

*Tel: 55 21 2562-7045/7047/2618/2558
www.po.ufrj.br/basilio/

*MailAddress:
COPPE/UFRJ
Caixa Postal 68507
CEP 21941-972 Rio de Janeiro,RJ
Brazil



You may leave the list at any time by sending the command

SIGNOFF allstat

to [log in to unmask], leaving the subject line blank.