Implementation a Various Types of Machine Learning Approaches for Biomedical Datasets based on Sickle Cell Disorder

This paper presents implementation a various kinds of machine learning models to classify the dataset of sickle cell patients. Artificial intelligence techniques have served to strengthen the medical field in solving its problems and providing rapid technical methods with high efficiency instead of traditional methods that can be subject to many problems in diagnosis and to determine the appropriate treatment. The main objective of this study to obtain a highly qualified classifier capable of determining the suitable dose of the SCD patients from 9 classes. Through examining the techniques used in our experiment based on performance evaluation metrics and making sure that each model performs. We applied numerous models of machine learning classifiers to examine the sickle cell dataset based on the performance evaluation metrics. The outcomes obtained from all classifiers, show that the Naïve Bayes Classifier obtained poor results compared to other classifiers. While Levenberg-Marquardt Neural Network during the training phase obtained the highest performance and accuracy of 0.935222, AUC 0.963889. The test phase obtained an accuracy of 0.846444, AUC 0.871889.