Anemia Prediction Based on Rule Classification
Sahar Jasim Mohammed, Amjad A. Ahmed, Arshed A. Ahmad, Mohammed Sami Mohammed
Abstract
Iron or vitamin lack on the human body can be avoided to prevent some types of diseases like anemia. Anemia causes severe health problems like pregnancy complications or even heart real issues if not a predicted at earlier stages. In this paper, data is collected of 539 participants with 10 related features to each one. Three based rule classification techniques are applied to design on a curate anemia prediction system: ZeroR, OneR and PART to extract relevant of anemia dataset associated to "If" and "Then" a procedure. Among applied techniques, PART provided 85% accuracy higher than ZeroR and OneR. These techniques provided a benchmark of other applied technique to clarify sufficient knowledge of anemia data rules.