應用資料探勘於產品推薦以支援行銷決策策略-以IC設計業為例

Abstract

面對產業輪動快速的時代，其商業競爭環境趨向複雜多變，高科技產業逐漸走向產品多樣化及客製化的需求。消費市場環境的改變演變為企業對時間的敏感度提高，傳統的行銷思維在競爭激烈的環境中，將面臨到很大的挑戰。企業必須在不同的市場上找到具有相同產品需求的客戶，若企業無法因應市場變化提供客戶所需的產品，則客戶將轉向與競爭者交易，然企業本身便逐漸喪失競爭力。 本研究以消費性IC設計業為例，應用資料探勘的方法，提出一個產品推薦模型；藉由了解客戶的特性與需要並使用EM分群演算法對客戶進行分群；此外，透過關聯規則將客戶的交易資料進行產品關聯分析，協助行銷人員了解產品與客戶之間的關聯，期望帶給行銷人員在行銷決策策略不同的思維，以提升行銷效益

Facing the industrial era moving ahead rapidly, the competitive environment of business becomes more complex and volatile. The high-tech industries are tending to diversified and customized. Swift changes of customer market spur enterprise to focus on improving time sensitivity. Keeping the traditional marketing thinking, enterprises will encounter a great challenge in the competitive environment. Enterprises could get the strategy to meet market change, if they could provide customized products; otherwise, customers would transfer to competitors and enterprises would gradually lose their competitive edge. In this study, we propose a product recommendation approach by using consumer IC design industry as a case study. The proposed approach is based on the concept of association rule-based recommendation. The EM clustering method is applied to cluster customers into groups of customers with similar characteristics and product needs. Moreover, association rule mining is applied to extract the associations of products for recommendations by analyzing customer transactions. The proposed product recommendation model could support marketing sales to obtain more clear decision strategy by realizing the relationship between products and customers. Consequently, enterprises can improve marketing effectiveness and earn value.