Prediction of demand for blood bank products according to blood groups by a data mining approach by using neural networks

A limited supply chain opportunity between blood donation and blood transfusion requires the optimization of each stage in the supply chain of the blood, and in particular, to achieve more effective blood supply management in blood centers and hospitals. Since previous research has not predicted blood demand by blood products and blood groups. So in this research, demand of blood blank products is predicated according to blood groups by using artificial neural networks (Case study: Zanjan blood Transfusion Network). 

A data set including information on all actual whole blood donations from 21 March 2014 to 21 September 2016 is obtained by Zanjan Blood Bank from the national donor database. .for this subject, Database of Zanjan blood Transfusion Network was used. The goal of this article is to use ANN model to evaluate trends in FFP, PLT and RBC with eight blood groups of A-, A+, AB-, AB+, B-, B+, O- and O+ demand and supply and to predict how these will develop over 30 months.

Findings show that the best model of artificial neural networks with different neurons in order to predict the demand for blood bank products by the blood groups in the transmission network of Zanjan province has two delays and five neurons in the hidden layer. Also, the results show that the error value is close to each other in all three blood products but has different values.

The results showed that this method is capable of predicting. Accordingly, in order to obtain more suitable models, future researchers are suggested to study the combination of artificial neural networks with meta-algorithms.