Analisis Path Modeling Segmentation Partial Least Square (PATHMOX-PLS) pada Gambaran Klinis Pasien HIV/AIDS

Hiariey, Arlene Henny (2018) Analisis Path Modeling Segmentation Partial Least Square (PATHMOX-PLS) pada Gambaran Klinis Pasien HIV/AIDS. Masters thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
1316201014-Master_Thesis.pdf - Accepted Version

Download (6MB) | Preview

Abstract

Structural Equation Modelling (SEM) merupakan metode analisis statistik multivariat dapat menjelaskan struktur hubungan yang relatif kompleks dengan melibatkan banyak variabel. SEM berbasis komponen atau varian disebut Partial Least Square (PLS), merupakan metode yang lebih fleksibel dan powerfull dalam pengujian dengan bebas asumsi distribusi, skala dari indikator beragam, dan jumlah sampel tidak harus besar. PLS-PATHMOX melakukan segmentasi terhadap hasil model persamaan struktural untuk mengatasi observed heterogeneity. Spesifik permasalahan dalam kasus HIV/AIDS dapat dilihat melalui gambaran klinis penderita HIV/AIDS, berdasarkan infeksi oportunistik, kadar CD4 dan kualitas hidup. Penelitian ini bertujuan untuk mendapatkan model persamaan struktural (SEM-PLS) gambaran klinis pasien HIV/AIDS dan menyusun segmentasi hasil model persamaan struktural gambaran klinis HIV/AIDS berdasarkan pendekatan PLS-PATHMOX. Penelitian ini menggunakan data sekunder dan berasal dari rekam medik 150 pasien HIV/AIDS di Puskemas Kabupaten Pasuruan tahun 2017 sejumlah 150 pasien. Hasil menunjukkan bahwa faktor predisposisi, faktor pemungkin dan faktor penguat signifikan terhadap gambaran klinis. Nilai R2 yang dihasilkan sebesar 0,8 sehingga dapat disimpulkan bahwa model mepunyai kemampuan yang baik dalam menjelaskan data. Pada hasil pengelompokan dengan menggunakan PATHMOX-PLS, gambaran klinis HIV/AIDS dikelompokkan oleh variabel segmentasi jenis kelamin dan tingkat pendidikan. Split node yang pertama signifikan dengan p-value 0,0032 dan split node yang kedua signifikan dengan p-value 0,0006. Gambaran klinis pasien HIV/AIDS dapat dimodelkan menggunakan PATHMOX-PLS dan lebih mudah untuk segmentasi menjadi segmen kecil, lebih spesifik sesuai dengan kesaaman. =============================================================================================== Structural Equation Modeling (SEM) is a method of multivariate statistical analysis that can describes the structure of a relatively complex relationships by engaging many of the variables. SEM-based on component or a variant called Partial Least Square (PLS), is a method that is more flexible and powerful in testing with free distribution assumptions, the scale of the indicator variety, and the number of samples must not be large. PLS-PATHMOX does a segmentation model equations results against structural to address the observed heterogeneity. Specific problems in the case of HIV/AIDS can be seen through clinical overview sufferers HIV/AIDS, based on opportunistic infections CD4 levels, and quality of life. This research aims to get Structural Equation Model (SEM-PLS) clinical picture of HIV/AIDS patients and develop structural equation model of segmentation results overview of HIV/AIDS clinical approach based on PLS-PATHMOX. Secondary data were gathered and derived from the medical records of 150 patients with HIV/AIDS at the health centers of the Pasuruan Regency in 2017. The results showed that the predisposing factors, enabling factors and strengthening factors influenced the clinical picture of HIV/AIDS patients. R2 produced by 0,8 so it can be conclude that model can explain data. Based on the grouping analysis using PATHMOX-PLS, the clinical picture of the HIV/AIDS patient was grouped by the segmentation variables i.e. gender (male and female) and patient’s education levels. First split the resulting significant with p-value segmentation used 0.0032 and second split the resulting significant with p-value segmentation used 0.0006. So it can be conclude the clinical picture of HIV/AIDS patients can be modeled using PATHMOX-PLS and it easier to take a large patient base into smaller, more specific segments according to their similarities.

Item Type: Thesis (Masters)
Uncontrolled Keywords: Gambaran klinis HIV/AIDS, Kabupaten Pasuruan, PLS-PATHMOX, Variabel Segmentasi
Subjects: Q Science > QA Mathematics
Q Science > QA Mathematics > QA278.3 Structural equation modeling.
Divisions: Faculty of Mathematics, Computation, and Data Science > Statistics > 49101-(S2) Master Thesis
Depositing User: Arlene Henny Hiariey
Date Deposited: 15 Jul 2021 09:17
Last Modified: 15 Jul 2021 09:17
URI: https://repository.its.ac.id/id/eprint/55606

Actions (login required)

View Item View Item