Titis Sriyanti (2015) Pemodelan Regresi Logistik Dan Multivariate Adaptive Regression Spline (Mars) Binary Response Untuk Prediksi Kejadian Hiv-Aids Di Kabupaten Banyuwangi. Thesis thesis, UNIVERSITAS AIRLANGGA.

	Text (HALAMAN JUDUL) 1. HALAMAN JUDUL .pdf Download (2MB)
	Text (ABSTRACT) 2. ABSTRACT .pdf Download (545kB)
	Text (DAFTAR ISI) 3. DAFTAR ISI .pdf Download (776kB)
	Text (BAB 1) 4. BAB I PENDAHULUAN .pdf Download (732kB)
	Text (BAB 2) 5. BAB II TINJAUAN PUSTAKA .pdf Restricted to Registered users only until 21 February 2023. Download (1MB) | Request a copy
	Text (BAB 3) 6. BAB III ALGORITMA PENELITIAN .pdf Restricted to Registered users only until 21 February 2023. Download (641kB) | Request a copy
	Text (BAB 4) 7. BAB IV METODE PENELITIAN .pdf Restricted to Registered users only until 21 February 2023. Download (739kB) | Request a copy
	Text (BAB 5) 8. BAB V HASIL DAN ANALISIS DATA .pdf Restricted to Registered users only until 21 February 2023. Download (1MB) | Request a copy
	Text (BAB 6) 9. BAB VI PEMBAHASAN .pdf Restricted to Registered users only until 21 February 2023. Download (752kB) | Request a copy
	Text (BAB 7) 10. BAB VII PENUTUP .pdf Restricted to Registered users only until 21 February 2023. Download (717kB) | Request a copy
	Text (DAFTAR PUSTAKA) 11. DAFTAR PUSTAKA .pdf Download (625kB)
	Text (LAMPIRAN) 12. LAMPIRAN .pdf Restricted to Registered users only until 21 February 2023. Download (1MB) | Request a copy

Abstract

One of regression analysis used to analyze binary response variable is logistic regression. Multivariate adaptive regression splines (MARS) is a nonparametric regression model that assumes the function of the relationship between predictor variables and response is unknown. The objective of this study is to obtain the best model between MARS Binary Response and Regression Logistic to predict the incidence of HIV-AIDS in Banyuwangi, and to see the predictor variables that affect incidence of HIV-AIDS in Banyuwangi. Study design used in this research is Non Reactive study. Population in this study is people with HIV-AIDS recorded in the regristration data from NGO Care HIV / AIDS Kelompok Kerja Bina Sehat Banyuwangi with a sample size of 261 respondents. Results of the data analysis with logistic regression in models get Y = 0.36 -2.902 X2 - 2.378 X3 + 4.439 X6 - 3,501 X7 influential variables were age, occupation, casual partners and condom use. Great value probability is 8.45. The best modeling with Binary Response MARS obtained the best model Y = 0844 + 0031 * BF1 - BF2 + 0044 * 0251 * 0323 * BF3 + BF7 - BF9 + 0033 * 0057 * 0024 * + BF11 BF12 BF13 + + 0426 * 0028 * BF15 - BF19 + 0510 * 0911 * BF21 - 0765 * BF23 - BF25 + 0352 * 0680 * BF27 Wherein the variables that affect the value of GCV in a row is casual partners, condom use, occupation, age, the pair remained, sex and marital status. From the second model used logistic regression models prediction accuracy is 90.8%, while the accuracy of the MARS models Binary Response of 97.3%. Therefore MARS Binary response models larger than the Logistic Regression Model, it can be concluded that the modeling results MARS Binary Response was better than the logistic regression modeling.

Actions (login required)

View Item