<![CDATA[Blood is an important fluid that naturally produced in the human body. When a human lost a lot of blood, a blood transfusion is needed . Blood for the transfusion is provided by a blood storage center in charge of estimating blood demand to minimize the excessive amount of blood in storage or wasted blood. Lack of blood supply can affect to the increased death of the patient, while an oversupply of blood until passes it shelf life (35 days) should also be avoided. In order to minimize the loss, a method to forecast the blood demand is needed, that is fuzzy time series. To increase the accuracy, the method is optimized with particle swarm optimization to determine the best interval in fuzzy time series. Based on the results of a series of tests, the optimum solution with average of cost value (MSE) of 60435,685 is obtained on 40 particles, 30 dimensions, 1.5 and 1.5 for the combination of and value respectively, the weight of inertia of 0.3, and the maximum number of iterations of 950. By using 12 testing data, the error rate generated by this system (MAPE) is 7.50330%.]]>
Copyrights © 2018
