CMStatistics 2020: Start Registration
View Submission - CMStatistics
Title: Non-crossing spatial autoregressive quantile model applied to dengue fever incident in Bandung, Indonesia Authors:  Yudhie Andriyana - Universitas Padjadjaran (Indonesia) [presenting]
Abstract: The existence of spatial dependence in a dataset needs to be accommodated by a proper model. One of the standard techniques often used is Spatial Autoregressive (SAR) model, which will be implemented to the incidence of dengue fever in Bandung, Indonesia. Dengue fever is an infectious disease that has impacts not only on a health aspect but also on social and economic aspects. Therefore, to prevent the disease, we need to control factors that influence the dengue's occurrence, considering the dependence between area. An interesting study is to know the model on the highest or the lowest risk of dengue fever, which cannot be solved by a classical regression technique. Therefore, we propose to use an existing technique called the quantile spatial autoregressive model. However, the technique is working with individual quantile objective function. In that case, it may lead to a crossings issue where the lower quantile levels may cross to the higher levels or vice versa. Hence, we propose a quantile technique to avoid such crossing problems.