Integration of Kansei engineering and artificial neural network toward the implementation of intelligent food packaging design based on consumer preferences

Packaging design innovation is one of the crucial strategies for consumer-oriented product development. Therefore, this research aimed to design intelligent food packaging (IFP) for beef products using an integrated approach of Kansei engineering (KE) and artificial neural network (ANN) based on consumer preferences. The results showed 37 valid and reliable Kansei words based on Kaiser-Meyer-Olkin measure (KMO), Bartlett's test of sphericity, and measure of sampling adequacy (MSA) using SPSS 26 software. Based on the results, the best ANN structure was achieved with the Traingd learning algorithm which had 418 inputs, 20 nodes in the hidden layer, and eight outputs with a training mean square error (MSE) of 0.0099991, a validation MSE of 0.0321, a training regression (R) of 0.99287, and a validation R of 0.98928. Therefore, the best IFP design for beef products based on consumer preferences could be achieved by integrating KE and ANN methods.