[ABE-L] Seminários de Estatística e Ciência de Dados do DEST-UFBA

['Lizandra Castilho Fabio' via abe-l@ime.usp.br]

Prezados, bom dia!



Na próxima quinta-feira, dia *06 de outubro, às 17h00*, daremos
continuidade aos Seminários em Estatística e Ciência de Dados do
Departamento de Estatística da Universidade Federal da Bahia.



This seminary will be presented by *Prof. **Olushina Olawale Awe*. He is an
Elected Member of the International Statistical Institute (ISI) and a
Fellow of the African Scientific Institute, USA. He is the First LISA
Fellow and presently the LISA 2020 Global Engagement Ambassador in the LISA
2020 Global Network of the University of Colorado, Boulder, USA.  He holds
a PhD in Statistics from the University of Ibadan, Nigeria and MBA from
Obafemi Awolowo University, Ile-Ife, Nigeria. He is an Affiliate member of
the African Academy of Sciences (AAS) and an immediate past Council Member
of the International Society for Business and Industrial Statistics (ISBIS)
(2017-2021).  Currently, he is visiting scholar at the Institute of
Mathematics, Statistics and Scientific Computing, University of Campinas,
Brazil (2021-2023).



Following some details about this seminary:

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

 *Title*: Modelling Point Referenced Spatial Count Data: A Poisson Process
Approach


Discriminant analysis is a standard statistical learning method for modern
data analysis. In many practical data applications, there are often a large
number of pre-processed heteroscedastic features. It is well known that the
Linear Discriminant Analysis (LDA) is quite suboptimal for the analysis of
high dimensional heteroscedastic data because of the inherent singularity
and instability of the within-class scatter matrix. However, shrinkage
discriminant analysis (SDA) and its variants often perform better due to
its robustness against inherent multicollinearity and heteroscedasticity.
In this work, we propose some newly modified discriminant classification
algorithms based on the SDA. We show an estimation consistency property of
three newly modified supervised methods, and compare with the
heteroscedastic discriminant algorithm and other existing competitors. Our
empirical application shows that the proposed algorithms perform moderately
well for datasets with high dimensions and unequal covariance structures
when applied to nutrition data. The sensitivity and specificity of the
target classes ranges from 70-100\%. The balanced accuracy of all the
algorithms ranges from 50 to 75\% for a three-class problem. It was
concluded that shrinkage discriminant algorithms perform well with high
sensitivity for classifying heteroscedastic survey data with high
dimensions.


*Dia:* 06 de outubro de 2022.



*Hora:* 17h00.

*Link: **https:// meet.google.com/rcp-qgtt-jwz
<http://meet.google.com/rcp-qgtt-jwz>*

-- 

*Lizandra Castilho Fabio*
Associate Professor
Department of Statistics
Federal University of Bahia, Brazil
-------------- Próxima Parte ----------
Um anexo em HTML foi limpo...
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20220928/ad8c1725/attachment-0001.htm>