Enhancing Blood Disease Diagnosis Using Artificial Neural Networks: A Predictive Model with Robust Performance

Blood diseases, ranging from anemia and leukemia to clotting disorders, are a major concern in the health sector because of the worldwide morbidity and mortality rates they cause. Diagnosis of the diseases in question is very important if treatment is to be effective and the outcome for the patient improved. However, current trends in machine learning, especially Artificial Neural Network (ANN) has great potential in improving the diagnostic works and processes by automating them. ANN learns structured blood sample metrics and figures out complex relationships that are necessary and important in classifying diseasesIn this research, we present a predictive model for blood disease classification using ANN. The dataset was cleaned and prepared through feature engineering strategies and included several other diseases metrics. The class weighting technique was employed to ensure that while the model learnt all the classes, the robust performance across the entire diseases category was guaranteed. The model got an accuracy of 95% in this model with a majority of the disease classes having high precision and recall.