Estimasi Fungsi Baseline Hazard Model Cox dengan Metode Maximum Penalized Likelihood pada Data Durasi Pencarian Kerja di DKI Jakarta

Yudhajaya, Agatha Sandira (2024) Estimasi Fungsi Baseline Hazard Model Cox dengan Metode Maximum Penalized Likelihood pada Data Durasi Pencarian Kerja di DKI Jakarta. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5006201038-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2026. Download (2MB) | Request a copy

Abstract

Tingkat pengangguran merupakan salah satu masalah penting yang diperhatikan pemerintah Indonesia. Pengangguran adalah keadaan di mana seseorang di usia produktifnya belum memiliki pekerjaan dan sedang berusaha untuk mencari pekerjaan. Tingkat pengangguran di Indonesia sangat berfluktuasi dari tahun ke tahun. Begitu pula di DKI Jakarta yang merupakan pusat perekonomian Indonesia. Pengangguran di DKI Jakarta termasuk cukup tinggi dibandingkan provinsi-provinsi lainnya. Pada penelitian ini, akan dilakukan analisis survival pada durasi pencarian kerja di DKI Jakarta dengan model Cox Proportional Hazard. Analisis faktor-faktor yang memengaruhi kecenderungan seseorang untuk mendapatkan pekerjaan dilakukan dengan dua metode: maximum partial likelihood dan maximum penalized likelihood di mana pada metode maximum penalized likelihood selain akan didapatkan koefisien regresi, juga akan didapatkan estimasi fungsi baseline hazard. Dengan menggunakan data durasi pencarian kerja dari Survei Ketenagakerjaan Nasional (SAKERNAS) Badan Pusat Statistik (BPS) pada Februari 2023 didapatkan hasil model Cox Proportional Hazard dengan kedua metode. Kedua model memberikan hasil bahwa variabel usia, pendidikan, dan pelatihan adalah faktor yang memiliki pengaruh signifikan. Pada metode maximum partial likelihood didapatkan hazard ratio variabel usia, pendidikan SD-SMA, pendidikan perguruan tinggi, pelatihan adalah sebesar 1,066512; 0,312891; 0,598126; dan 1,536562 secara berturut-turut. Sedangkan dengan metode maximum penalized likelihood didapatkan hazard ratio variabel usia, pendidikan SD-SMA, pendidikan perguruan tinggi, pelatihan adalah sebesar 1,082; 0,144; 0,320; dan 1,800 secara berturut-turut. Estimasi fungsi baseline hazard pada metode maximum penalized likelihood ini tertinggi pada bulan pertama dan terus menurun hingga bulan ke-18 selanjutnya baseline hazard relatif konstan pada bulan ke-19 hingga 25. Didapatkan bahwa metode maximum penalized likelihood lebih baik dibandingkan metode maximum partial likelihood dengan nilai AIC 1001,61 dibandingkan 1702,17.
=================================================================================================================================
Unemployment rate is one of the main problem in Indonesia which need government full attention. Unemployment is a situation where an individual in their productive age have no job and is currently looking for a job. Unemployment rate in Indonesia continue to change every year, including DKI Jakarta. As the center of economics in Indonesia, DKI Jakarta has a quite high unemployment rate compared to other province. In this research, survival analysis will be performed to analyze the job search duration in in DKI Jakarta with Cox Proportional Hazard model. The analysis will use two methods: maximum partial likelihood dan maximum penalized likelihood. With maximum penalized likelihood method, not only factors affecting the job search duration and regression coefficient estimation will be obtained, but also baseline hazard function estimation will also be performed. Using job search duration in Jakarta data from National Workforce Survey by BPS in February 2023, two Cox Proportional Hazard models are obtained. Both model result in three significant variable which are age, education, and training. Model with maximum partial likelihood method obtained hazard ratio for age, education (elementary to high school), edulation (college/university), and training of 1,066512; 0,312891; 0,598126; and 1,536562 consecutively. While model with maximum penalized likelihood method obtained hazard ratio for age, education (elementary to high school), edulation (college/university), and training of 1,082; 0,144; 0,320; and 1,800 consecutively. The maximum penalized likelihood baseline hazard function estimates the highest point is at the first month and gradually decreases to the lowest point at 18th month. The next months (19th – 25th month) baseline hazard function is relatively constant. It is found that the maximum penalized likelihood method results in better estimation compared to maximum partial likelihood with AIC score 1001,61 versus 1702,17.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Analisis Survival, Cox Proportional Hazard, Durasi Pencarian Kerja, Maximum Partial Likelihood, Maximum Penalized Likelihood, Job Search Duration, Survival Analysis
Subjects:	Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Agatha Sandira Yudhajaya
Date Deposited:	17 Jan 2024 06:50
Last Modified:	17 Jan 2024 06:50
URI:	http://repository.its.ac.id/id/eprint/105531

Actions (login required)

View Item