在探討影響消費者行為意向的研究領域,因果模式為研究者最常使用的方法。而線性結構關係(LISREL)與類神經網路便為近來研究者較常使用的分析方法。為了解這兩種方法在客運旅客行為意向應用之差異,本研究針對此二種方法進行實證研究,並討論其在理論上與應用時的特性。本研究以國道客運台中-高雄線、台北-台南線之旅客為調查對象。先利用LISREL驗證所建構之結構方程式模式之適配性,再以NN針對全連式與非全連式兩種模式進行分析。結果顯示,對於已知因果關係的模式,LISREL是一種方便且有效的分析工具;而NN則不論是否事先推導出因果關係,經由適當的學習程序後仍有相當的預測力。最後則提出本研究之結論與建議。
Casual model is one of the most powerful instruments in consumer behavioral intentions research. LISREL and neural networks have recently become the popular method to perform casual model. To recognize the differences between LISREL and neural networks methods, this study applied these two methods to passengers’ behavioral intentions and discussed the characteristics in theory and practice. We took passengers of Taichung-Kaohsiung line and Taipei-Tainan line of intercity bus on national freeway as our samples. First, this study applied LISREL to test the goodness of fit of research model. Then two kinds of NN model were tested which one is full connected network and the other is non-full connection network. The results indicated that LISREL can be a convenience and effective analysis tool while the causal relationships were known. On the other hand, no matter the causality was derived out in advance, NN still has suitable prediction after the proper training procedure. At last, we proposed our conclusions and suggestions based on our study results.