Multivariate Data Analysis to Classify Blood Donors Utilizing Supervised Learning Models

A safe and adequate supply of blood and blood components can be achieved by recruiting, retaining, and encouraging donor populations, thereby ensuring the availability of blood needed for transfusions around the clock and throughout the year. Blood donation services can be improved by technology improvements that aid in the analysis of potential donors, as well as the development of applications to assess donor health questionnaires. The work conducted in the paper classifies blood donors with the status Eligible or Deferred using synthetic donor data comprising data fields created based on eligibility criteria for blood donation. The health vitals of the donors from among the data fields emerge as decision-makers in classifying the donor as Eligible or Deferred. Multivariate data analysis and modeling using supervised machine learning models help relate synthetic data to real-time data and draw useful patterns and conclusions. Building trust in the classified results obtained using machine learning models is facilitated by the use of Explainable Artificial Intelligence (XAI) tools. Further work can be done on analyzing and modeling data of different available blood groups, age groups, and gender which can be deciding factors in recruiting and targeting potential donors to build efficient blood donation services.