Nooraeni, Rani
Pusat Penelitian dan Pengabdian Masyarakat Politeknik Statistika STIS

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

KAJIAN EFEK SPASIAL KASUS DIFTERI DENGAN GEOGRAPHICALLY WEIGHTED NEGATIVE BINOMIAL REGRESSION (GWNBR) Mustika, Diva Arum; Nooraeni, Rani; IJSA, Indonesian Journal of Statistics and Its Applications
Indonesian Journal of Statistics and Applications Vol 3 No 1 (2019)
Publisher : Departemen Statistika, IPB dengan Forum Perguruan Tinggi Statistika (FORSTAT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29244/ijsa.v3i1.185

Abstract

Diphtheria is an infectious disease caused by the Corynebacterium diphtheriae bacteria. Indonesia is the country with the most cases of diphtheria in Southeast Asia and ranks third in the world. In 2016, cases of diphtheria increased by 65 percent and became Extraordinary Events (KLB) in Indonesia, even though during 2013 to 2015 the number of cases of diphtheria has decreased. The province that has the highest number of diphtheria cases in Indonesia in 2016 is East Java. Diphtheria is centered and spread in certain districts / cities in East Java Province so that there are indications of spatial effects in the spread of diphtheria. Because data on the number of diphtheria cases overdispersed and indicated spatial effects in its spread, the main method used in this study was Geographically Weighted Negative Binomial Regression (GWNBR). This method will be compared with other alternative methods namely Poisson regression method and Negative Binomial Regression to get the best modeling. Based on the AIC value of each model it can be concluded that the best method for modeling the number of diphtheria cases is GWNBR. The modeling results with GWNBR show that there is indeed a spatial influence on the number of diphtheria cases and risk factors in East Java Province in 2016. The percentage of DPT-HB3 / DPT-HB-Hib3 immunization coverage is not significant in all observation areas, while the percentage of drug and vaccine availability is significant at entire observation area.
Metode Cluster Menggunakan Kombinasi Algoritma Cluster K-Prototype dan Algoritma Genetika untuk Data Bertipe Campuran Nooraeni, Rani
Jurnal Aplikasi Statistika & Komputasi Statistik Vol 7 No 2 (2015): Journal of Statistical Aplication and Statistical Computing
Publisher : Pusat Penelitian dan Pengabdian Masyarakat Politeknik Statistika STIS

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (895.38 KB) | DOI: 10.34123/jurnalasks.v7i2.23

Abstract

Clustering adalah salah metode utama pada data mining yang berguna untuk mengeksplorasi data. Membagi suatu data set berukuran besar ke dalam cluster yang sehomogen mungkin adalah tujuan dalam metode data mining. Salah satu metode clustering konvensional yaitu algoritma K-Means efisien untuk dataset berukuran besar dan tipe data numerik tapi tidak untuk data kategorikal. Algoritma K-Prototype menghilangkan keterbatasan pada data numerik tapi dapat juga digunakan pada data kategorikal. Namun solusi yang dihasilkan oleh kedua algoritma tersebut merupakan solusi lokal optimal dimana salah satu penyebabnya adalah penentuan pusat cluster awal. Untuk menghadapi masalah tersebut maka algoritma genetika menjadi salah satu usulan yang dapat digunakan untuk mengoptimalkan hasil pengclusteran dengan K-Prototype. Hasil dari penelitian menunjukkan optimasi pusat cluster dengan algoritma genetika berhasil meningkatkan akurasi hasil cluster dengan K-Prototype.