層級貝氏模型應用於消費者之社群與網路購物行為分析之研究

The Relevance of Consumer Social Behavior and Size of Purchase by Using Hierarchical Bayesian Analysis

陸昀泓(Yun-Hung Lu)

指導教授 ： 任立中

國立臺灣大學/管理學院/國際企業學研究所/碩士(2014年)

https://doi.org/10.6342/NTU.2014.00809

摘要

本研究嘗試分析消費者於社群媒體平台上的行為與購買金額大小之關聯性，將每個消費者視為獨立且具有異質性的個體，並運用消費者過去的歷史交易資料與粉絲團上的行為資料分析出個別消費者的特定社群與購買行為模式。研究主旨在於深入探究消費者的社群行為(Social Behavior)與其電子商務上購買金額之影響性，因此透過5項自變數以量化消費者的社群行為，並且使用層級貝式統計分析方法來探究與購買金額之關聯性。
　　本研究運用國內某家販售母嬰用品之電子商務廠商的資料庫，資料庫分成三種，包括電子商務的交易資料庫、粉絲團上社群行為資料以及透過安裝於電腦上的Cookie以追蹤消費者的平時上網的時間點(digital footprint)，利用以上資料做為分析。在變數的選取上，本研究以消費者於電子商務中的購買金額為依變數，而自變數則分別為：消費者對粉絲團上Marketing類型文章按讚次數、消費者對粉絲團上Communication類型文章按讚次數、消費者於粉絲團上中性情緒評論次數、消費者於粉絲團上負面情緒評論次數、消費者於粉絲團上正面情緒評論次數。再將消費者平時上網的時間點(digital footprint)透過K means集群分析進行分群，然後將分群資料做為第二層自變數，並運用層級貝式統計分析以探究以下兩個主題：(1)消費者的購買金額會受到哪些社群行為影響？以及(2)消費者在購買金額與各個自變數之行為模式是否在不同的上網時間的分群變數上有顯著差異？
　　本研究或許可提供企業與行銷人員一個資料庫加上社群行銷之分析架構，但在消費者行為之探索上各個網站未必相同，因此行銷人員可依循本研究之方法以了解消費者之社群行為。

關鍵字

資料庫行銷 ； 購買金額預測 ； 社群行為 ； 層級貝式統計 ； 網路足跡

並列摘要

This study attempts to analyze the behavior of consumers in the size of purchase and its relevance of online social behavior, each consumers is individual and heterogeneity, and by the past history of the transaction data with the facebook fan page data to analyze the particular pattern of one individual consumer its social behavior and also purchase behavior. Through five independent variables to quantify the social behavior, and use hierarchical Bayesian statistical analysis to explore the following themes.
　　This study acquire the database from the e-retailer in Taiwan, which sells baby and mother products. The database is divided into three types, including e-commerce transaction database, the behavior of the facebook fans page and the time consumer surf on the internet through the Cookie installed on consumer’s computer on, also known as digital footprint. This study attempts to put the size of purchase as dependent variable, while the independent variables are: the numbers of "likes" on the marketing articles, the article of communication, the comment of neutral emotion, the comment of negative emotion and the comment of positive emotion. Information on category variables of consumer surfing internet and then set as the second independent variable, and use the hierarchical Bayesian statistical analysis to explore the following two themes: (1) Will the size of purchase be affected by the social behavior? And (2) whether every beta coefficient is significant differences in category variables?
　　However, this study may provide marketing personnel a framework for database and social marketing analysis, consumer behavior in the exploration of various sites may not be the same, so they can follow the method of this study to understand consumer social and purchase behavior.

並列關鍵字

database marketing ； sales predictions ； social behavior ； Hierarchical Bayesian analysis ； digital footprint

參考文獻

3. 李來錫、謝明晃(2006)，社群信任、產品涉入與認知風險對於虛擬社群購買動機之影響研究，行銷評論，第3卷，第2期，頁149-168。

1. Chang, A. M., Kannan, P. K., & Whinston, A. B. (1999), “Consumers extent of evaluation in brand choices”, The Journal of Business, Chicago, 72(2), 229-251.

2. Chiu, C. M., Hsu, M., & Wang, M. H. (2006), “Understanding knowledge sharing in virtual communities: An integration of social capital and social cognitive theories”, Decision Support Systems, 42(3), 1872-1889.

3. Dellarocas, C., X. M. Zhang and N. F. Awad (2007), “Exploring the Value of Online Product Reviews in Forecasting Sales: the Case of Motion Pictures”, Journal of Interactive Marketing, Vol. 21, No. 4, pp. 23-45.

4. Judith A. Chevalier and Dina Mayzlin (2006), “The Effect of Word of Mouth on Sales: Online Book Reviews”, Journal of Marketing Research Vol. XLIII, pp.345–354.