WITHDRAWN: Anemiascreening in pregnant women by using feature selection with data classification algorithms

Abstract

Abstract Anemia in pregnancy is a significant issue that risks the life of the mother and the fetus, if not given proper care. Though it is very common in women during pregnancy, a better understanding of the outcome is very important for the development of the treatment strategies. Data mining techniques led to better performance in decision making. This work aims to introduce various feature selection methods and classification algorithms based on the classes. They are compared to finding out the best possible method for accurate prediction of the existence of anemia. The anemia prediction in women during pregnancy took many categories including employment, type of residence - urban and rural for the different types of anemia classes namely, Mild, Severe or moderate and Non-anemic. This paper discussing on anemia prediction in three phases namely (i) Gaussnominal classification; (ii) VectNeighbour Classification and (iii) Random Prediction (PR) classification with appropriate feature selection techniques are described. The experimental result of phase 3 Random Prediction (RP) classification achieved 98.94% that was more effective than the previous works like Gaussnomial and VectNeighbour classifications. The final result showed that the new model outperformed the other models inaccuracy.