标题：KLASIFIKASI KEMISKINAN DI KOTA SEMARANG MENGGUNAKAN ALGORITMA CHISQUARE AUTOMATIC INTERACTION DETECTION (CHAID) DAN CLASSIFICATION AND REGRESSION TREE (CART)
摘要：Decreasing poverty level is the first goal of
Sustainable Development Goals (SDGs). Poverty in
Central Java from 2002 to 2017 has decreased, as well as
the city of Semarang. Therefore, it is necessary to
examine the factors that determine the decline in poverty
classification in the city of Semarang. The classification
analysis in statistics uses one classification tree. Several
methods using classification trees include CART,
CHAID, C45 and ID3 algorithms. In this study the
methods used were CART and CHAID Algorithms.
CART and CHAID algorithms are binary classification
trees. The CART separation rules use split goodness op,
while CHAID uses CHI-Square. In the analysis, the value
of using CART was 95.2% while CHAID was 95.2%.
While the factors that influence poverty classification
using CHAID include the acceptance of poor rice, the
main building materials of the house walls, and the main
fuel for cooking. Whereas with the CART Algorithm the
variables that influence are the main fuels for cooking,
poor rice receipts, the number of household members,
final disposal sites, sources of drinking water, the
household head's business field, roofing materials, and
building walls.
