Application FCM in modelling DIR for Selangor using negative binomial GAM

• Main Authors: Mohamad, Nazeera, Che Him, Norziha, Rusiman, Mohd Saifullah, Sufahani, Suliadi, Muhammad Jamil, Siti Afiqah
• Format: Article
• Published: Science Publishing Corporation 2018
• Subjects: TA Engineering (General). Civil engineering (General); QA299.6-433 Analysis; TA329-348 Engineering mathematics. Engineering analysis
• Online Access: http://eprints.uthm.edu.my/5465/

This study attempts to obtain the best fitted model among two clusters which describe the relationship between dengue incidence rate (DIR) and relevant covariates such as climatic and non-climatic variables. The significant variables include amount of rainfall and number of rainy days with lag 0 until 3 months, number of locality and population density. Fuzzy C-Means clustering (FCM) was applied in clustering DIR data based on the value of membership function. The boundary of membership function has been set as 0.5. There are two clusters identified in this study with Cluster 1 consist of 569 data and Cluster 2 consist of 43 data. Then, this study developed models to predict future dengue incidences in Selangor by using negative binomial Generalised Additive Model (GAM). The result shows that the model able to be one of tools for future development in controlling and reducing the number of dengue cases particularly in Selangor, Malaysia as well as other states.