Annisa Ratna Dewi, 080513384 (2010) Estimasi Model Regresi Nonparametrik Aditif Dengan Respon Berdistribusi Poisson Berdasarkan Estimator Kernel. Skripsi thesis, UNIVERSITAS AIRLANGGA.
|
Text (Abstrak)
gdlhub-gdl-s1-2011-dewiannisa-15072-abstrak-e.pdf Download (229kB) | Preview |
|
Text (Full Text)
gdlhub-gdl-s1-2011-dewiannisa-13079-kkckkm-e.pdf Restricted to Registered users only Download (1MB) | Request a copy |
Abstract
Regresi Poisson merupakan suatu metode statistik yang digunakan untuk menghubungkan variabel respon Y yang berdistribusi Poisson dengan variabel prediktor X . Distribusi Poisson berasal dari distribusi keluarga eksponensial dengan variabel prediktor yang dihubungkan secara linear melalui fungsi penghubung, dinamakan Generalized Linear Model (GLM). Generalized Additive Model (GAM) merupakan perluasan atau generalisasi dari GLM dengan menggantikan bentuk linear dengan bentuk aditif. Fungsi ( ) . j m merupakan fungsi smooth yang akan diestimasi menggunakan estimator Kernel. Untuk mengestimasi model regresi nonparametrik aditif dengan respon berdistribusi Poisson berdasarkan estimator kernel digunakan algoritma local scoring , dengan membuat program pada software R. Data yang digunakan dalam penerapan yaitu data jumlah kematian ibu maternal menurut puskesmas Kabupaten Cirebon tahun 2006. Variabel respon yang digunakan adalah jumlah kematian ibu maternal. Variabel prediktornya adalah persentase ibu hamil yang mendapatkan pelayanan imunisasi TT2 ( tetanus toxoid ) dan jumlah puskesmas pembantu dan keliling (kendaraan roda empat). Dari hasil analisa data diperoleh bahwa setiap kenaikan 1% ibu hamil yang mendapatkan pelayanan imunisasi TT2 dengan asumsi jumlah puskesmas pembantu dan keliling tetap, akan menurunkan jumlah kematian ibu maternal secara rata-rata sebesar 0,98 kali sebelumnya dan setiap penambahan sebuah puskesmas pembantu atau keliling, dengan asumsi persentase ibu hamil yang mendapatkan pelayanan imunisasi TT2 tetap, akan menurunkan jumlah kematian ibu maternal secara rata-rata sebesar 0,97 kali sebelumnya Kata Kunci : GAM, Regresi Poisson, Esti mator Kernel, Algoritma Local Scoring. ii ADLN Perpustakaan Universitas Airlangga Skripsi Estimasi Model Regresi Nonparametrik Aditif Dengan Respon Berdistribusi . . . Ratna Dewi, Annisa Annisa Ratna Dewi, 2010. Estimation of Additive Nonparametric Regression Model by Using Poisson Distributed Response Based on Kernel Estimator. This Skripsi guided by Nur Chamidah, S.Si, M.Si and Toha Saifudin, S.Si, M.Si. Mathematics, Faculty of Science and Technology, Airlangga University. ABSTRACT Poisson regression is a statistic method which is used to relate Y response variable which have Poisson distribution to X predictor variables. The response of Poisson distribution is assumed to have a distribution from exponential family with predictor variable that is connected by linear through the link function, is called Generalized Linear Model (GLM). Generalized Additive Model is an extension or generalizes form from GLM substitutes linear model to additive model. Function () . j m is a smooth function that will be estimated using Kernel estimator. For estimating a Poisson additive nonparametric regression model based on kernel estimator, we use local scoring algorithm, by making program in R software. We applied the model to maternal mortality data according to the public health centre in Cirebon Regency in 2006. Response variable which is used is the number of maternal mortality and the predictor variables are percentage of pregnancy that have gotten second tetanus toxoid immunization service and the number of mobile car for public health centre ( PUSKESMAS ). Based on the data are obtained that every 1% increasing of pregnancy that have gotten second tetanus toxoid immunization service, with assumption of total health centre are fix, will decrease the number of the maternal mortality averagely as much as 0,98 times before and every increasing a mobile car for public health centre ( PUSKESMAS ), with assumption percentage of pregnancy that have gotten second tetanus toxoid immunization service are fix, will decrease the number of the maternal mortality averagely as much as 0,97 times before. Keywords : GAM, Poisson Regression, Kernel Estimator, Local Scoring Algorithm.
Item Type: | Thesis (Skripsi) | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
Additional Information: | KKC KK MPM 75-10 Dew e | |||||||||
Uncontrolled Keywords: | REGRESSION ANALYSIS; NONPARAMETRIC STATISTIC | |||||||||
Subjects: | Q Science > QA Mathematics > QA276-280 Mathematical Analysis | |||||||||
Divisions: | 08. Fakultas Sains dan Teknologi > Matematika | |||||||||
Creators: |
|
|||||||||
Contributors: |
|
|||||||||
Depositing User: | Tn Septian Eko Budianto | |||||||||
Date Deposited: | 07 Mar 2011 12:00 | |||||||||
Last Modified: | 19 Jul 2016 03:58 | |||||||||
URI: | http://repository.unair.ac.id/id/eprint/25442 | |||||||||
Sosial Share: | ||||||||||
Actions (login required)
View Item |