- 學位論文
運用層級貝氏方法建構消費反應模式
Constructing Consume Response Model By Hierarchical Bayesian Methodology
摘要
綜觀過去對消費者行為之研究，基本上皆著重在分析人口統計變數及生活型態變數對消費者購買決策之影響，但隨著行銷環境的改變，顧客的品味及生活背景皆已不同，其所反應給企業的將是不同的顧客需求。因此企業如何了解不同顧客之異質性並辨識出顧客之差異，已成為企業進行行銷活動時最重要的關鍵之一。然而傳統學者在利用最近購買日(Recency)、購買頻率(Frequency)及購買金額(Monetary) (即RFM模式)或其他方法進行顧客價值分析時，皆將顧客視為是一個均質性的市場；事實上，此種假設已經不符合現在顧客的現況，因此近年來，學者除了開始將顧客異質性納入分析消費反應模式之外，也發展出許多的分析方法，在許\\\\\\\\多的發展方法裡，層級貝氏方法是其中最常被學者使用的技術之一。理論上用於傳統的RFM模式是以各別維度的分析來識別顧客的價值，因此對於同時考量兩種以上維度來解釋消費者行為的部份較少著墨。在本論文中我們嘗試提出一消費反應模式：在同時考量購買頻率及購買金額兩種維度的情境下來解釋消費者行為。在模式建構的過程中，我們採用了層級貝氏方法與蒙地卡羅馬可夫鏈( Markov Chain Monte Carlo, MCMC)模擬技術來進行模式中參數的估計；然後，透過估計的結果來解讀消費者的行為。
並列摘要
Recent research on customer behavior is often focused either on demographic variables or on life-style variables. Moreover, most of these research directly adopted RFM (Recency , Frequency and Monetary) technique to define what they so called customer values. However, faced with the marketing trend of an individualization- and differentiation-oriented market, businesses should have been deeming the issue of “how to incorporate consumers’ heterogeneous preferences with variables of consumers’ demographics” as their top priority. In this research, a brand new response model of consumers was proposed to analyze the customer values. In order to investigate the consumers’ heterogeneity, a hierarchical Bayesian model was developed by considering individual customer purchase amount and consuming counts with negative Binormial distribution. To demonstrate the effectiveness of the proposed model, the data from one of the Taiwan’s famous supermarkets was utilized. Based on the application results, we found that customer purchase amount and consuming counts can be successfully identified and predicted.