<![CDATA[Blood supply is a crucial aspect for UTD which must meet the demand for blood for those who need it. UTD Surabaya City faces challenges in meeting blood needs caused by the uncertainty of blood demand which varies and is individualized according to the recipient's clinical condition which has an impact on the quality of UTD Surabaya City services, thus creating challenges in meeting blood needs optimally. Therefore, it is necessary to predict blood demand to assist UTD Surabaya City in ensuring adequate blood stock, planning the blood stock needs that will be requested, and avoiding stock overstocks and stock shortages. To overcome this, blood demand is predicted using the Autogressive Moving Average (ARIMA) and Adaptive Neuro Fuzzy Inference System (ANFIS) approaches. This combination of the ARIMA-ANFIS method combines the advantages of ARIMA in capturing linear patterns and ANFIS in handling non-linear patterns from ARIMA residuals. The prediction results from the ANFIS model will be added to the prediction results from the ARIMA model to obtain a hybrid ARIMA-ANFIS model. The ARIMA-ANFIS model is used to predict the number of blood requests by combining ARIMA predictions and residuals modeled using ANFIS. This process includes stationarity analysis, selecting the best ARIMA model, residual modeling with ANFIS, as well as performance evaluation using MAPE to ensure prediction accuracy. The best ARIMA (6,1,0) model was obtained with the lowest AIC value of -153.838, then from the ARIMA modeling results the residuals were obtained as input for ANFIS modeling. Analysis shows that the ARIMA-ANFIS hybrid model has better performance, with a MAPE value of 5.28%, compared to the ARIMA model which only achieved a MAPE of 6.21%.]]>
Copyrights © 2025
