An Intelligent Blood Management System Using OPTICS Clustering and Real-Time Donor Tracking

Effective blood donor selection involves the efficient prioritization of donors, compatibility checking, inventory validation, and spatial location-based donor identification. This research proposes an intelligent blood management system including four-class donor priority classification, eligibility and blood group validation, inventory first validation, spatial clustering using OPTICS algorithm, geographic prioritization, and donor tracking after acceptance. For donor priority classification, Blood Transfusion Service Center dataset comprising 748 donor data and four numerical attributes-recency, frequency, monetary value of blood volume, and time-has been used. The developed classification algorithm shows 0.95 precision, 0.93 recall, and 0.94 F1-score while the confusion matrix analysis shows that there is an effective segregation between the four classes of donor priorities having 91.37% classification accuracy. Furthermore, the results of multiclass ROC analysis confirm the discriminatory capability of the model having AUC values ranging from 0.95 to 0.99.