Skripsi

ANALISIS KLASIFIKASI TINGKAT KEMISKINAN DENGAN ALGORITMA XGBOOST: STUDI KASUS KABUPATEN BANYUASIN

---

Poverty remains a critical issue in Banyuasin Regency, South Sumatra, with a poverty rate of 9.31% in 2024, exceeding the national average of 8.47%. This study develops a poverty classification model using the XGBoost algorithm with SUSENAS 2024 data from Banyuasin Regency. The research employs 16 socioeconomic variables as features, with poverty status determined based on the Banyuasin poverty line of Rp539,283 per capita per month. To address severe class imbalance (ratio 1:25), SMOTE was applied to the training data. The model was evaluated on 141 test samples maintaining original distribution. Results demonstrate excellent performance with 97.16% accuracy and 99.56% AUC-ROC. For the minority class (Poor), the model achieved 80% recall, successfully detecting 4 out of 5 poor households, with 57.14% precision and 66.67% F1-score. Feature importance analysis revealed household expenditure, number of household members, and head of household education as most influential factors. The model shows significant potential as a screening tool for social assistance targeting, though field verification remains necessary. This research contributes to poverty classification methodology in Indonesia by demonstrating the effectiveness of combining advanced machine learning with data balancing techniques for highly imbalanced social data.

---

Availability

Detail Information

Series Title

-

Call Number

T1923102026

Publisher

Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2026

Collation

xvi, 155 hlm.; ilus.; tab.; 29 cm.

Language

Indonesia

ISBN/ISSN

-

Classification

339.460 7

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Prodi Teknik Informatika
Analisis Tingkat Kemiskinan

Specific Detail Info

-

Statement of Responsibility

MI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment