Open AccessPENERAPAN FEATURE WEIGHTING OPTIMIZED PADA NAÏVE BAYES UNTUK PREDIKSI PROSES PERSALINAN
Author(s) -
Hilda Amalia,
Achmad Baroqah Pohan,
Siti Masripah
Publication year - 2019
Publication title -
pilar nusa mandiri/pilar nusa mandiri
Language(s) - English
Resource type - Journals
eISSN - 2527-6514
pISSN - 1978-1946
DOI - 10.33480/pilar.v15i1.3
Subject(s) - weighting , naive bayes classifier , bayes' theorem , medicine , computer science , artificial intelligence , bayesian probability , support vector machine , radiology
Birth of a baby is something that is very desirable for every married couple. All parties expect safety for mothers and babies who have just been born. Medical personnel make various efforts to help the delivery process run smoothly and the mother and baby survive. But in the labor process not all the baby's birth process runs smoothly. Problems often occur during labor. There are several obstacles so that there is a risk of labor, namely maternal and infant mortality. Every mother wants to be able to give birth to a baby normally, but due to medical reasons the delivery process is done by cesarean. The act of choosing a type of delivery faster can affect the safety of the mother and baby. The selection of the cesarean method is carried out late so it will increase the risk of maternal and infant mortality. For this reason, it is necessary to conduct research by using labor delivery data so that they can choose the right type of labor. In this study the classification of maternity labor will be carried out with data mining methods, namely Naive Bayes, which are improved by using the Optimize Weight (PSO) method. Naive Bayes was able to produce a high accuracy value for processing labor data for mothers, namely 94%. The final results of this study obtained the value of naïve bayes performance that can be improved by the Optimize Weights (PSO) method to be better at 98%