%0 Thesis
%9 Skripsi
%A Shun Nafingatun Nurul Ummah, NIM.: 22106010028
%B FAKULTAS SAINS DAN TEKNOLOGI
%D 2026
%F digilib:78914
%I UIN SUNAN KALIJAGA YOGYAKARTA
%K Regresi logistik LASSO; AIC; stunting; SSGI
%P 117
%T OPTIMISASI REGRESI LOGISTIK LASSO DENGAN KRITERIA AKAIKE INFORMATION CRITERION (AIC) UNTUK KLASIFIKASI FAKTOR RISIKO STUNTING PADA ANAK BERDASARKAN DATA SSGI 2024
%U https://digilib.uin-suka.ac.id/id/eprint/78914/
%X In binary logistic regression, multicollinearity among predictor variables can cause parameter estimates to become unstable, thereby reducing the quality of the model. To address this issue, LASSO (Least Absolute Shrinkage and Selection Operator) logistic regression is used, which applies an L1 norm penalty, thereby enabling simultaneous variable selection and coefficient shrinkage. The penalty parameter λ was determined using the Akaike Information Criterion (AIC) to obtain a model that strikes a balance between model fit and complexity. This method was applied to data from the 2024 Indonesian Nutrition Status Survey (SSGI), which covers 38 provinces in Indonesia, to identify factors influencing stunting status. The results of the study show that LASSO logistic regression using the AIC criterion produces a simpler model with seven predictor variables: Interpregnancy Interval, Low BirthWeight (LBW), Exclusive Breastfeeding, Minimum Dietary Diversity (MDD), Complete Immunization, Acute Respiratory Infections (ARI), and High Maternal Education. The model has a lower AIC value than full logistic regression, with an AUC of 1.00 and a classification accuracy of 100%. These results indicate that the LASSO model using the AIC criterion is capable of producing a parsimonious model with excellent classification performance.
%Z Sri Utami Zuliana, S.Si., M.Sc., Ph.D.