透過您的圖書館登入
IP:3.145.183.137
  • 學位論文

整合反應時間之層級貝氏聯合分析模型

Integrated Hierarchical Bayesian Conjoint Model with Response Time

指導教授 : 任立中

摘要


「市場分析」為行銷的第一步,行銷人員依據市場分析為基礎,制定行銷策略,再依照行銷策略發展策劃行銷活動,諸如Kotler之行銷4P、Schultz之整合行銷傳播IMC等,因此市場分析著實為企業行銷能否成功的主要關鍵的第一步。既然市場分析的重要性不言可喻,因此吾人當須追求市場分析的「正確性」,以求建立更穩固的行銷基礎。分析正確性仰賴於分析工具的選擇及資料質量的優劣,聯合分析是利用來分析屬性偏好的統計工具,被廣泛於應用各個領域,在商業上適合被用於分析消費者的偏好結構,例如產品、廣告、通路、定價之屬性偏好等,藉由正確地瞭解消費者的偏好結構,才能制定正確的行銷策略。 本研究主要即針對聯合分析的進階方法─適應性聯合分析法(adaptive conjoint analysis),進行模型的改良,透過結合受測者答題反應時間的資料,建立聯合分析迴歸資料的潛藏資料擴充模型(latent data augmentation)。潛藏資料擴充模型乃為依據反應時間特性,以電腦模擬抽樣的方式擴充產生個別受測者的潛藏觀察值資料,改善問卷資料的質與量。 本研究同時利用先進的層級貝氏統計方法(hierarchical Bayesian statistics)來估計聯合分析資料的結果。透過建構一個層級貝氏迴歸模型,並利用電腦進行馬可夫鍊蒙地卡羅模擬方法(Markov Chain-Monte Carlo),模擬預估個別受測者的屬性偏好值。 此外,本研究以兩種研究產品之問卷資料作模型之實證分析,並同時與傳統應用於處理聯合分析資料的統計模型進行比較,透過模型的實證分析資料,探討不同模型的預測能力(good of fit)。

並列摘要


Market research is the first stage of marketing, and then comes to the marketing strategy, and finally comes to the marketing activities. Since market research is so fundamental to marketing, the success of a marketing plan will depend heavily on the accuracy of market research. There are two ways to improve the accuracy of market research. One is to choose the appropriate data analysis tool, and the other one is to improve the quality and the quantity of data. Conjoint analysis is one of the best tools to analyze consumers’ preference of attributes, and is widely applied in many other fields of researches. Adaptive conjoint analysis, the advanced statistical software of conjoint, is used in the study to collect the respondents’ conjoint data and response time data efficiently with the help of computer technology. A latent data augmentation model is proposed in this study, which uses response time data to expand the latent conjoint data depending on the features of response time, and is used to improve the quality and quantity of raw data. Bayesian statistics is introduced in the study to constitute of the hierarchical Bayesian regression model to estimate each respondent’s preference. Two actual survey data of conjoint analysis are conducted in the proposed model to exam the effectiveness of prediction. Besides, this study drew a comparison of good of fit and effectiveness of prediction among the proposed model and other three statistical models, which are widely and traditionally used in conducting conjoint analysis.

參考文獻


葉書芳(2005),多屬性決策模式下之消費者偏好分析。國立台灣大學國際企業學研究所碩士論文。
謝文瑋(2005),整合聯合分析與反應時間在顧客偏好預測之應用。國立台灣大學國際企業學研究所碩士論文。
陳靜怡(2005),購買量與購買時程雙變量之預測─層級貝氏潛藏行為模型之建構。國立台灣大學國際企業學研究所博士論文。
Allenby, Greg M., Neeraj Arora, and James L. Ginter (1995), “Incorporating Prior Knowledge into the Analysis of Conjoint Studies”, Journal of Marketing Research, Vol.32, May, 152-162
Allenby, Greg M., Neeraj Arora, and James L. Ginter (1998), “On the Heterogeneity of Demand”, Journal of Marketing Research, Vol.35, Aug., 384-389

延伸閱讀