雙重通路複合式協同過濾之產品推薦研究

Abstract

在網際網路便捷的現代社會，企業逐漸朝向多重通路(Multi-Channel)行銷模式發展，顧客能選擇自己較喜好的通路來採購商品。企業若能整合與分析各通路的交易記錄，便能為顧客提供個人化的推薦服務，並提升顧客忠誠度。 為降低交易記錄稀疏問題(Data Sparsity)，本研究提出一個雙重通路複合式協同過濾(CF)推薦方法，整合顧客在不同通路下的交易記錄，並基於Nearst-Neighbor的協同過濾推薦方法，結合User-based CF與Item-based CF進行產品推薦，以提升推薦品質。實驗結果顯示，本研究所提之雙重通路複合式協同過濾法能提升產品推薦品質。

The convenience of Internet makes enterprises employ Multi-Channel to develop the marketing models. Customers can choose prefer channels to make purchases from enterprises. Integrating and analyzing transactions from Multi-Channel can provide personal recommendation to increase customer loyalty. To reduce the issue of data sparsity in transaction records, this research proposes a Hybrid Collaborative Filtering (CF) in dual channels for product recommendation. The proposed method employs Nearst-Neighbor approach by integrating the User-based CF and Item-based CF. The experimental results show that the proposed hybrid approach in dual channels can improve recommendation quality.