Combining Lisrel And Bayesian Network To Predict Tourism Loyalty

This study proposes an analytic approach that combines LISREL and Bayesian networks (BN) to examine factors influencing tourism loyalty and predict a tourist's loyalty level. LISREL is used to verify the hypothesized relationships proposed in the research model. Subsequently, the supported relationships are used as the BN network structure for prediction. 452 valid samples were collected from tourists with the tour experience of the Toyugi hot spring resort, Taiwan. Compared with other prediction methods, our approach yielded better results than those of back-propagation neural networks (BPN) or classification and regression trees (CART) for 10-fold cross-validation.