Thalassemia Prediction Using a Machine Learning Approach

Abstract

Thalassemia is one kind of genetic blood disease/disorder which is caused if human body can‟t produce sufficient hemoglobin. It is known that hemoglobin is a very common essential part of anyone‟s body. RBC‟s of human body don‟t work efficiently if there is any lacking of hemoglobin. Then a little amount of healthy RBC‟s travel in one‟s bloodstream. Oxygen which is carried by red blood cell is kind of food, that food cells can utilize to work. Due to lacking of sufficient healthy RBC‟s, sufficient oxygen can‟t be delivered to every cells of the body, which can be a reason to cause a person to anemia, that is responsible to damage organs and lead one to death. In this research we are working about predicting the existence of Thalassemia with ML, an important part of AI. We implemented very popular ML algorithms on our processed dataset. We used k-nearest neighbor (kNN), logistic regression, support vector machine (SVM), naïve Bayes, random forest, adaptive boosting (ADA boosting), XGBoost, decision tree, multilayer perception (MLP) and gradient boosting classifier. In our work, out of ten algorithms, ADA BOOST algorithm gave the greatest output which is related on accuracy and it was 100%.