MODELING OF RARE DISEASE FREQUENCY DATA USING ZERO INFLATED POISSON (ZIP) AUTOREGRESSION WITH LOCATION EFFECT
<p align="justify">In health and medical practice, it is interesting to know frequency of the disease at a location observed over time. For diseases with rare events, there will be many zeros in data. Modelling disease occurrence is important to predict probability and frequency of o...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/30707 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | <p align="justify">In health and medical practice, it is interesting to know frequency of the disease at a location observed over time. For diseases with rare events, there will be many zeros in data. Modelling disease occurrence is important to predict probability and frequency of occurrence for next time. As for modeling count data commonly used Poisson model. However, variance of data will be greater than mean, and Poisson model is no longer appropriate. One alternative that can be used is Zero Inflated Poisson (ZIP) distribusion. Generalized Linear Model (GLM) is used to construct ZIP Autoregression model that depend on frequency of previous times. Neighborhood effect is added to regressor to know effect frequency of occurence in surrounding neighbors to frequency of occurrence at the location to be modeled. Applications of this model are used in Syphilis disease frequency data in the State of Nevada, Oregon, and California. Effect of neighborhood is added to regressor with uniform and squared inverse distance weighted.<p align="justify"> |
|---|