Pengembangan Model Spasial Multivariate Adaptive Generalized Poisson Regression Splines (Studi Kasus : Jumlah Kejadian Tuberkulosis)

Yasmirullah, Septia Devi Prihastuti (2024) Pengembangan Model Spasial Multivariate Adaptive Generalized Poisson Regression Splines (Studi Kasus : Jumlah Kejadian Tuberkulosis). Doctoral thesis, Institut Teknologi Sepuluh Nopember.

Text 06211960010009-Dissertation.pdf - Accepted Version Restricted to Repository staff only until 1 April 2026. Download (9MB) | Request a copy

Abstract

Multivariate Adaptive Regression Splines (MARS) merupakan salah satu metode regresi nonparametrik yang dapat mengakomodasi efek aditif dan efek interaksi antar variabel prediktor. Secara umum, MARS telah digunakan untuk memodelkan pasangan data dengan respon numerik atau kategorik. Salah satu jenis data numerik yang perlu mendapat perhatian khusus dalam pemodelan adalah data jenis cacahan (count). Data count sering dijumpai terutama di bidang kesehatan. Adanya data count mendorong berkembangnya teori dan penerapan metode MARS, yaitu Multivariate Adaptive Generalized Poisson Regression Splines (MAGPRS). MAGPRS merupakan kombinasi dari MARS dan Generalized Poisson Regression. Tetapi keterlibatan faktor dependensi spasial dalam analisis data, menyebabkan keterbatasan model MAGPRS dalam melakukan analisis data. Oleh sebab itu dibutuhkan suatu metode yang dapat mengakomodasi faktor dependensi spasial. Hal ini juga memotivasi pengembangan teori dan penerapan metode spasial dan MAGPRS, sehingga dapat menjadi salah satu alternatif untuk menyelesaikan adanya efek dependensi spasial. Penerapan model spasial MAGPRS akan diaplikasikan pada data jumlah Tuberkulosis (TB).
=================================================================================================================================
Multivariate Adaptive Regression Splines (MARS) is a nonparametric regression method that can accommodate additive and interaction effects among predictor variables. MARS is generally used to model pairs of data with numerical or categorical responses. One type of numerical data that needs special attention in modeling is count data. Count data is frequently observed, particularly in the health sector. The existence of count data motivates the development of the theory and application of the MARS method, which is the Multivariate Adaptive Generalized Poisson Regression Splines (MAGPRS). The MAGPRS is a combination of MARS and generalized Poisson regression. Nevertheless, the involvement of the spatial dependency factor in data analysis causes the limitations of the MAGPRS model in data analysis. Hence, we require a technique that can effectively incorporate the spatial dependency factor. Additionally, it motivates the development of theory and the application of spatial and MAGPRS models, making it a viable alternative for addressing the spatial dependency effect. The spatial MAGPRS model will be applied to the number of tuberculosis (TB) incidence.

Item Type:	Thesis (Doctoral)
Uncontrolled Keywords:	Data Count, Spasial, Count Data, Multivariate Adaptive Regression Spline (MARS), Multivariate Adaptive Generalized Poisson Regression Spline (MAGPRS), Spatial
Subjects:	Q Science > QA Mathematics > QA401 Mathematical models.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49001-(S3) PhD Thesis
Depositing User:	Septia Devi Prihastuti Yasmirullah
Date Deposited:	04 Mar 2024 02:27
Last Modified:	04 Mar 2024 02:27
URI:	http://repository.its.ac.id/id/eprint/107757

Actions (login required)

View Item