Analisis Kurva Roc Pada Model Logit Dalam Pemodelan Determinan Lansia Bekerja Di Kawasan Timur Indonesia

Abstract

Binary logistic regression is used for probability modeling or to predict binary response variables (Success / Failure) from one or more explanatory variables that are continuous or categorical. In carrying out this analysis, there are several ways to test the suitability of the resulting model, and one of them is the area under the ROC curve. The application of the analysis method in this study is the determinant of the elderly population to work. The population of the elderly in Indonesia is increasing every year. Many views that the elderly depend on other residents, especially in terms of the economy. However, if seen from the percentage of elderly working in Indonesia, it is increasing, including the elderly in KTI. The purpose of this study is to determine the characteristics of the elderly in KTI, know the factors that influence the decision of the elderly population to work in KTI and find out the tendency of variables that affect the decision of the elderly to work in KTI. The data used are raw data from Badan Pusat Statistik (BPS) was Survei Sosial Ekonomi Nasional (Susenas) Kor March 2018. This study using descriptive analysis methods and binary logistic regression. The results are that the variables that significantly influence the decisions of the elderly to work are residence, gender, age, education, family status, marital status, health complaints, and health insurance. Elderly who has characteristics residing in rural, male sex, classified as young elderly (60-69 years old), has the highest level of elementary school education, has the status of KRT in his family, is married, has no complaints health, and not having health insurance will have a greater tendency to decide to work.

Pharmacophore Modeling and Binding Affinity of Secondary Metabolites from Angelica keiskei to HMG Co-A Reductase

“…The area under the ROC curve (Figure 3) was in the range of zero to one or 0-100%. The higher the area under the curve, the better the prediction of the model for active compounds than the decoy [32].…”

“…of zero to one or 0-100%. The higher the area under the curve, the better the prediction of the model for active compounds than the decoy [32]. Model 7 recognized 970 hits with an AUC value of 0.8%, an EF value of 6.4, and a GH score of 0.3.…”

“…Pharmacophore validation was carried out using 1 structure-based and 10 ligand-based pharmacophore models. The higher the area under the curve, the better the prediction of the model for active compounds than decoy [32]. The best model, namely ligand-based model 7, was sensitive in identifying 201 active compounds out of 209, representing 67% of the total active compounds.…”

Erosion Prediction and Soil and Water Conservation Scenario Using The SWAT Model

“…The ROC curve is the result of plotting true signal (sensitivity) and false signal (1-specificity) across the entire range of possible cutoff points. The range of the area under the ROC curve is from zero to one (Nur & Oktora, 2020). This curve demonstrates the probability or accuracy level of the model.…”

Karakteristik Pengemudi Sepeda Motor Dalam Model Peluang Kecelakaan (Studi Kasus: Perlintasan Sebidang di Jalur Perlintasan Langsung (JPL) 325 Lamongan)

“…Dinyatakan valid jika Ho = μo ≥ ttabel maka signifikan (valid) dan jika H1 = μ1 ≤ ttabel maka tidak signifikan (tidak valid). Sedangkan dinyatakan reliabel jika nilai Cronbach Alpha (α) > 0,70 [27]. Jika suatu instrumen dinyatakan tidak valid dan tidak reliabel, maka perlu dilakukan perbaikan dan diujikan kembali kepada responden.…”