Modeling and comparison of bonding strength of impregnated wood material by using different methods: Artificial neural network and multiple linear regression
In this study, the effects of vacuum time, diffusion time and pressing time on the bonding strength of Larix decidua wood impregnated with Immersol-Aqua and bonded with Klebit-303 were investigated. The vacuum time, diffusion time, and pressing time were predicted by using the artificial neural network (ANN) model and multiple linear regression (MLR) methods and the results of ANN and MLR methods were compared. The highest bonding strength (7.664 N. mm-2) was achieved when the vacuum time, the diffusion time and the pressing time were 20, 60 and 60 minutes, respectively, while the lowest value (4.62 N. mm-2) was achieved when the vacuum time, the diffusion time and the pressing time were 80, 120 and 20 minutes, respectively. The model results are as follows: The MAPE value for testing phase in the ANN was 7.266 and R2 value was 0.751 whereas the MAPE value of the MLR was 9.365 and R2 value was 0.558. The ANN model has been found to have better prediction performance than the MLR model.