15746

1619 دكتري

دكتري

آمار

اصفهان : دانشگاه صنعتي اصفهان

1399

هفدهم، 126ص. : مصور، جدول، نمودار

1399/06/22

كتابنامه

رياضي

رياضي

1399/06/25

2631515

Abstract Spatial scan statistic has been widely employed in spatial disease surveillance and spatialcluster detection However the over dispersion and excess of zeros are often presented in real world data causing not only the violation of likelihood assumption for the Poisson model but also excessive Type I error or false alarms In this study we propose the Bell scan andthe zero inflated Bell scan statistics which cover the over dispersion and or excess of zerosin the data The proposed scan methods can be potentially applied to the event data in asimple way Considering zero inflated models we compare the Bell Poisson and binomialscan statistics based on relative risk bias precision recall of cluster detection and power Byour simulations we show that the Bell scan is a robust and powerful alternative in comparisonwith the traditional scan models We finally illustrate the new methodology with two realdata scan analyses On the other hand the spatial scan statistics based on the Poisson and binomial modelsrely on Monte Carlo simulation and they are time consuming to scan big maps Hence wepropose some algorithms to detect irregular shape clusters using Poisson binomial and Bellmodels Then we apply these algorithms on big maps By simulation we show that the irreg ular Bell scan is robust comparing classical models in the detection of non circular clusters Finally we find spatial clusters on a medical image Keywords Spatial scan statistics Count data Minimum spanning three Validity index Bell distribution

صفيه محمودي

ماركوس پراتس، فردي كاستلارس

عادل محمد پور، ريحانه ريخته گران، مجيد جعفري خالدي