Analyzing and Predicting Anaemia with Advanced Machine Learning Techniques with Comparative Analysis

The rising prevalence of anaemia poses significant health challenges globally, necessitating accurate and timely diagnostic methods.This work applies the AdaBoost classification model to a Complete Blood Count (CBC) dataset to provide a reliable machine learning method for anaemia prediction.The methodology used consists of extensive preprocessing including data cleaning, one-hot encoding, z score normalization and automated feature selection to preserve the data's integrity and make the model simpler.The experimental results proved that the presented AdaBoost model achieved a promising accuracy of 92.7%, good precision of 84%, fair recall of 94% and good F1 value of 88.7%, indicating effectively well-balanced classification performance.Further, ROC curve analysis, with AUC of 0.94, confirms superior discriminatory capability.It compared their results to existing models, LogitBoost and Random Forest (RF); LogitBoost yields an accuracy slightly low at 89.3% and RF at 67.1%.The results highlight the capabilities of the AdaBoost model for early, accurate anaemia detection, providing substantial improvements over conventional diagnostic measures and improving clinical decision making in real time healthcare applications.