以消費者購買決策階段為基礎之適性化推薦系統設計與績效之研究

A Study for the Design and Performance of the Adaptive Re-commender System-Based on the Consumer's Pur-chase-Decision Stage

吳肇銘(Chao-Ming Wu)

《資訊管理展望》 10卷2期 (2008/12) Pp. 19-43

https://doi.org/10.29715/JIMCSA.200812.0002

摘要

隨著網際網路上各式各樣的網站、資訊激增，「資訊超載」巳成為網友搜尋資訊時一項嚴重的問題。Ricci(2002)提出電子商務型網站可利用「推薦系統」來建議商品或提供資訊，以降低消費者「資訊超載」的問題，並有效協助其制定購物決策。然而目前許多推薦系統，由於並未考量消費者所處之購買決策階段不同，所需之資訊類型應有所不同，所以在進行推薦時往往無法符合消費者之需要。因此，本研究特別將消費者「購買決策階段」之概念納入推薦系統，提出一「適性化推薦系統」之架構，並實際設計出系一統。而為驗證此「適性化推薦系統」之績效，本研究透過一實地實驗，經資料分析後，獲得以下結果：（一）推薦系統中加入「消費者購買決策階段判斷」，有助於推薦系統之績效；（二）採用不同類型之推薦演算方法時，推薦系統之推薦成效會有所不同；（三）「消費者購買決策階段判斷方法」與「推薦方法」之交互作用會對推薦系統績效產生影響。

關鍵字

推薦系統 ； 適性化 ； 消費者決策過程 ； 網路行銷

並列摘要

With the rapid growth of many websites, ”information overloading” is a critical problem for web users in information search. Ricci (2002) argued that EC website should suggest products or provide information to customer by ”recommender system” to reduce the ”information overloading” problem and to make effective purchase-decision. However, many recommender systems do not take account of customer's different purchase-decision stage. They provide the information which is not fit for customer's need. This study takes ”customer's purchase-decision stage” into account to design and implement an ”adaptive recommender system based on the consumer's purchase-decision stage”. In order to evaluate the performance of this adaptive recommender system, a field experiment was executed. The main findings are in the following: (1) the recommender system with the judgment of the customer's purchase-decision stage has better performance; (2) the recommender systems using different kinds of algorithms for recommendation have different performances; (3) the interaction recommender system.

並列關鍵字

customer decision process ； recommender system ； adaptation ； internet marketing