Skripsi
CLASSIFICATION AND REGRESSION TREE UNTUK MENGKLASIFIKASIKAN TINGKAT PREVALENSI KETIDAKCUKUPAN KONSUMSI PANGAN PADA KABUPATEN/KOTA DI INDONESIA
This study aims to classify the Prevalence of Undernourishment (PoU) in regencies/cities in Indonesia. The method used is Classification and Regression Tree (CART). The data used is secondary data covering Rice Production (Produksi Beras, PB), Population Density (Kepadatan Penduduk, KP), Percentage of Poor Population (Persentase Penduduk Miskin, PPM), Percentage of Per Capita Food Expenditure (Persentase Pengeluaran per Kapita Makanan, PPKM), Desirable Dietary Pattern (Pola Pangan Harapan, PPH), Human Development Index (Indeks Pembangunan Manusia, IPM), and Food Security Index (Indeks Ketahanan Pangan, IKP). The analysis results show that the initial classification tree involves six variables, while pruning the classification tree results in five variables, namely IKP, PPM, PPKM, PPH, and IPM. The classification performance evaluation resulted in an accuracy of 84.42% with a kappa coefficient of 45.26%, which is classified as moderate. The sensitivity value of 85.53%, a precision of 81.91%, and an F1-score of 83.55% indicate a high classification ability. Analysis of the level of importance of the variables shows that the IKP made the highest contribution (43.16%), followed by the PPM (18.56%), PPKM (10.44%), PPH (10.30%), IPM (9.58%), KP (7.60%), and PB (0.36%) in the formation of the PoU level classification tree. Overall, the CART method was able to produce a good classification of the PoU level in regencies/cities in Indonesia. Keyword : Prevalence of Undernourishment, Classification and Regression Tree, Classification
No other version available