Data to Diagnosis Machine Learning Models for Accurate Anemia Classification

Anemia is a common hematologic condition that requires prompt and precise diagnosis in order to be managed and treated. This study investigates the use of different machine learning models to categorize different kinds of anemia based on complete blood count (CBC) data. To determine the best method for diagnosing anemia, we tested a number of models, such as XGBoost, Random Forest Classifier, and Decision Tree Classifier. The collection included CBC data from several medical facilities that had been tagged with anemia diagnosis. After thorough data pretreatment, features were chosen utilizing techniques like ensemble model feature significance, Variance Inflation Factor (VIF), and Predictive Power Score (PPS). GridSearchCV was used to tune the models' hyperparameters while 5-fold cross-validation was used for training and evaluation. The Decision Tree Classifier outperformed more intricate ensemble techniques, achieving the greatest balanced accuracy score of 94.17%, according to the results. Confusion matrices demonstrated its accuracy and recall while validating its strong performance. The work emphasizes how straightforward decision tree models can be used for medical diagnosis tasks, especially when datasets are properly preprocessed.