Feature Selection Based Classification of Blood Cells using Machine Learning Models

Blood also called Red Gold is considered to be so valuable for its purpose like transferring oxygen and nutrients to body parts and removing unwanted waste. Another role of blood is the immunity it provides along with clotting of blood during injuries. The main ingredients of the blood are the blood cells classified into three major types of erythrocytes, leukocytes, and thrombocytes. This research study focuses on the classification of leukocytes into five types as mentioned above with a preprocessing step of removing contiguous regions of the cell. Next, feature extraction is done by using standard methods and feature selection is done in three stages for enhancing the accuracy. The final step is classification, it is performed by using machine learning classifiers and all the results are compared based on the accuracy. When compared with all the models used, the SVM model provides an accuracy of 96.84% in classifying the cell types.