Pengembangan Model Prediksi Laju Pengeringan Pada Irisan Wortel (Daucus Carota) Berbasis Regresi Linier Berganda (Rlb) Dan Jaringan Syaraf Tiruan (Jst)

Abstract

Drying is a process of reducing water content as a result of heat transfer and water mass. Carrots (Daucus carota) are one of the ingredients that are often dried for additional instant food serving and their drying rate is an essential part to be studied. This research conducted to analyze the effect of thickness variations of carrot slices and developed prediction models for dry rate. The samples were fresh carrot slices at a thickness of 2 mm, 4 mm, and 6 mm with a water content of 90.72%. The measurement time was every 0.5 hours for 5.5 hours of drying with three replications. The one-way ANOVA test and DMRT resulted in a significant difference of each carrot slices thickness to dry rate with a significance level of 99%. Two prediction models were developed, namely the Multiple Linear Regression model (MLR) and the Artificial Neural Network model (ANN). Training and validation of the RLB model resulted in RMSE values of 10.622 and 10.409, then R2 values of 0.66 and 0.64. While training and validation of the ANN model resulted in RMSE values of 1.237 and 2.099 then R2 values of 0.996 and 0.992