Abstract Details
Activity Number:
|
129
|
Type:
|
Contributed
|
Date/Time:
|
Monday, August 10, 2015 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Statistics in Epidemiology
|
Abstract #317681
|
View Presentation
|
Title:
|
Using IRT Models to Estimate and Visualize Spatial Clusters
|
Author(s):
|
Andre Cancado* and Antonio Eduardo Gomes and Cibele Queiroz da Silva and Fernando Luiz Pereira Oliveira and Luiz Duczmal
|
Companies:
|
University of Brasilia and University of Brasilia and University of Brasilia and Federal University of Ouro Preto and Universidade Federal de Minas Gerais
|
Keywords:
|
spatial disease clusters ;
IRT
|
Abstract:
|
The most widely used version of the spatial scan statistic is the circular scan. However, circular windows are not always adequate to correctly describe the true solution when the cluster has a irregular shape. Furthermore, this methodology does not provide any measure of the relevance of each region to the most likely cluster. Based on a recently proposed tool, the intensity function, we build a more accurate way of defining the uncertainty in the delineation of spatial clusters using Item Response Theory (IRT) models for adjusting the probability that each region belongs to the anomaly. IRT is used in situations in which there are latent characteristics of interest - such as the proficiency in a given area or the extent of depression of a given individual - and we use observed variables to measure how much of the latent trait an individual possesses. We propose the use of the 2PL model to predict the probability that a given region of the map belongs to the real cluster. We use a bootstrap framework to obtain the necessary data and fit the model. Our tests show that the IRT approach is able to accurately pinpoint the true cluster. Simulations and applications will be presented.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2015 program
|
For program information, contact the JSM Registration Department or phone (888) 231-3473.
For Professional Development information, contact the Education Department.
The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
2015 JSM Online Program Home
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.