PREDICTION OF ANEMIA USING THE PARTICLE SWARM OPTIMIZATION (PSO) AND NAÏVE BAYES ALGORITHM

Abstract

Purpose: Anemia is a nutritional disorder that is still often found in Indonesia. The main risk factors for iron deficiency anemia are low iron intake, poor iron absorption, and periods of life when the need for iron is high such as during growth, pregnancy, and breastfeeding. Anemia can generally occur in pregnant women, teenagers, the elderly and even babies who have anemia.

Methods/Study design/approach: This research uses the Naive Bayes and PSO algorithms, and the dataset used comes from the kaggel.com Anemia dataset. The number of data records is 1421 data consisting of 5 attributes and 1 label. This data set is used to predict whether a patient is likely to suffer from anemia.

Result/Findings: Based on the results of testing the Naïve Bayes and PSO algorithm models which were carried out through confusion matrix evaluation, it was proven that the tests carried out by the Naïve Bayes algorithm were 93.88% and the tests carried out with Naïve Bayes and PSO had a high accuracy value, namely 94.02%.

Novelty/Originality/Value: The purpose of selecting information acquisition features is to select features or attributes that significantly influence anemia.

Keywords: Prediction, Anemia, Naive Bayes, Particle Swarm Optimization (PSO)