以層級貝氏模型探討於預算限制下序列決策模型之需求價格彈性推論

Abstract

過往在選擇式聯合分析中，實驗設計多半為觀測受試者之單一購買決策行為， 忽略市場消費者在挑選序列產品組合中預算遞減之動態性捕捉。本文旨在研究每 一位消費者對於同樣的菜單產品組合內容、不同的順序呈現方式，存在著不一樣 的價格敏感度，此次研究透過設計線上動態式調整菜單，並且在固定預算限制提 供給受試者的條件下，觀察受試者在面臨不同菜單順序之價格敏感度差異，利用 聯合分析法的架構，使用層級貝式迴歸分析、馬可夫鏈蒙地卡羅估計方式，以獲 得每一位獨立的受試者在各點餐階段的價格敏感度之變化，並做市場消費者分群， 以利於廠商在設計菜單時，不論是要針對每一位獨立的消費者作出客製化的菜單 推薦、或是將市場上消費者進行分群後並設計出數個版本的菜單架構，鎖定消費 者的動態價格敏感度來調整菜單順序的設計，有效的提高餐飲業者之利潤。

Choice-based Conjoint Analysis was mostly to observe the single purchase decision making in the past, ignoring the dynamically decreasing budget of consumers when selecting the sequential product combinations. This thesis aims to study the different price sensitivity of each consumer to the same menu product combination content but not the same presentation order. In this paper, we design an online dynamically adjustment menu, and given the fixed budget for subjects, to observe the price sensitivity when they take some menu of different presentation order, by using the framework of Conjoint Analysis, Hierarchical Bayesian regression analysis, and Markov chain Monte Carlo estimation to obtain everyone’s price sensitivities in each stage of the menu and do consumer clustering. So that when caterers design menus, whether it is to make customized menu recommendations for each independent consumer, or to group consumers in the market and design several versions menu, adjusting the menu according to the dynamic price sensitivity of consumers can effectively improve the profit of the catering industry.