Model Multiple Period Logit Menggunakan Pendekatan Maximum Likelihood (ML) Dan Bayesian Pada Data Penderita Kanker Payudara Di PT. C-Tech Lab Edwar Teknologi Tangerang

Soraya, Nazmi (2021) Model Multiple Period Logit Menggunakan Pendekatan Maximum Likelihood (ML) Dan Bayesian Pada Data Penderita Kanker Payudara Di PT. C-Tech Lab Edwar Teknologi Tangerang. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 06211850010023-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2023. Download (2MB) | Request a copy

Abstract

Analisis survival merupakan metode statistika yang berhubungan dengan
waktu, mulai dari waktu asal sampai terjadinya suatu peristiwa khusus (event).
Pada kasus tertentu suatu objek memiliki kondisi event yang dapat berubah-ubah
setiap waktu. Analisis survival yang dapat mendeteksi setiap perubahanperubahan
waktu tersebut adalah analisis multiple period logit. Multiple period
logit merupakan model logit yang diestimasi menggunakan data survival dengan
pengamatan antar objek bersifat independen. Fungsi likelihood dari multiple
period logit equvalent dengan fungsi likelihood biner, sehingga parameter model
multiple period logit dapat diestimasi menggunakan cara yang sama seperti model
regresi logistik biner biasa. Estimasi yang digunakan dalam penelitian ini adalah
estimasi klasik Maximum Likelihood (ML) dan estimasi Bayesian dengan prior
uniform dan multivaraiat normal.
Pada tahun 2020 terdapat 213.546 kasus kanker di Indonesia, kasus kanker
payudara meningkat menjadi 16,6% dengan kematian 9,6%. Saat ini telah ada alat
alternatif yang dianggap mampu menurunkan angka kematian pasien kanker
payudara, yaitu Electro Capacitive Cancer Therapy (ECCT). ECCT adalah alat
terapi berupa rompi yang di dalamnya terdapat medan listrik arus bolak-balik
dengan intensitas rendah (<30Vpp) dan frekuensi rendah (<100KHz) untuk
menghambat pertumbuhan sel kanker. Dari hasil analisis yang didapatkan pasien
kanker payudara yang mengalami metastasis, kondisi klinis, efek samping dan
lama penggunaan alat ECCT merupakan faktor yang mempengaruhi kematian
pasien kanker payudara yang menjalani terapi ECCT di PT. C-Tech Lab Edwar
Teknologi Tangerang pada tahun 2013-2017. Performa kedua estimasi tersebut
dilihat dari nilai C-index terbesar yaitu sebesar 62,4% dari estimasi Bayesian
dengan prior multivariat normal.
======================================================================================================
Survival analysis is a statistical method related to time, starting from the
time of origin to the occurrence of a special event. In certain cases, an object has
an event condition that can change at any time. Survival analysis that can detect
any changes in time is multiple period logit analysis. Multiple period logits is a
logit model which is estimated using survival data with independent observations
between objects. The likelihood function of the multiple period logits is
equivalent to the binary likelihood function so that the multiple period logit model
parameters can be estimated using the same way as the ordinary binary logistic
regression model. The estimates that will be used in this research are the calcic
Maximum Likelihood (ML) estimation and the Bayesian estimation with uniform
priors and normal multivariate.
In 2020 there were 213,546 cancer cases in Indonesia, breast cancer cases
increased to 16.6% with 9.6% mortality. Currently, there is an alternative tool that
is considered capable of reducing the mortality rate of breast cancer patients,
namely Electro Capacitive Cancer Therapy (ECCT). ECCT is a therapeutic tool in
the form of a vest in which there is an alternating current electric field with low
intensity (<30Vpp) and low frequency (<100KHz) to inhibit the growth of cancer
cells. From the results of the analysis, it was found that breast cancer patients who
experienced metastases, clinical conditions, side effects, and duration of use of the
ECCT device were factors that influenced the death of breast cancer patients who
underwent ECCT therapy at the PT. C-Tech Lab Edwar Teknologi Tangerang in
2013-2017. The performance of the two estimates is seen from the largest c-index
value, which is 62,4% of the Bayesian estimate with normal multivariate priors.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Kata kunci: Analisis Survival, Multiple Period Logit, Maximum Likelihood, Bayesian, Kanker Payudara, dan ECCT. Keywords: Survival Analysis, Multiple Period Logit, Maximum Likelihood, Bayesian, Breast Cancer, and ECCT
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) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Nazmi Soraya
Date Deposited:	09 Sep 2021 08:47
Last Modified:	09 Sep 2021 08:47
URI:	http://repository.its.ac.id/id/eprint/91922

Actions (login required)

View Item