文化大學機構典藏 CCUR:Item 987654321/44898
English  |  正體中文  |  简体中文  |  全文笔数/总笔数 : 46962/50828 (92%)
造访人次 : 12424671      在线人数 : 1103
RC Version 6.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
搜寻范围 查询小技巧:
  • 您可在西文检索词汇前后加上"双引号",以获取较精准的检索结果
  • 若欲以作者姓名搜寻,建议至进阶搜寻限定作者字段,可获得较完整数据
  • 进阶搜寻
    主页登入上传说明关于CCUR管理 到手机版


    jsp.display-item.identifier=請使用永久網址來引用或連結此文件: https://irlib.pccu.edu.tw/handle/987654321/44898


    题名: 運用協同過濾技術於最佳型顧客之個人化書籍推薦
    Using Collaborative Filtering for Personalized Book Recommendation for the Best Customers
    作者: 蔡佳芳
    贡献者: 資訊管理學系碩士在職專班
    关键词: 協同過濾
    顧客價值矩陣
    資料探勘
    混合式推薦
    Collaborative Filtering
    Customer Value Matrix
    Data Mining
    Hybrid Recommendation System
    日期: 2019
    上传时间: 2019-08-14 11:04:53 (UTC+8)
    摘要: 數位經濟的來臨加速了電子商務的發展,新購物平台不斷問世,除了開發新客源,如何了解並留住既有顧客,成為企業的關鍵課題。本研究對象為國內某網路購物平台,該平台主要銷售各式書籍,而書籍類商品並非民生必需品,一般都是有特定需求才會少量購買。而對喜歡閱讀的消費者來說,其購買次數及金額往往超出多數消費者的數倍。如果能讓網站能更精確的找出這些喜歡閱讀的最佳型顧客,並為其規劃專屬優惠行銷活動及個人化商品推薦,不但能提升商品銷量,也能減少行銷成本的支出。
    本研究首先對既有顧客進行消費分析先以RFM作為分群基礎,並使用Cascade K-means進行顧客分群,再參考顧客價值矩陣將分群結果區分為四群找出最佳型顧客,接者將以使用者為基礎及以項目為基礎所取得的最佳預測分數來進行不同權重的計算,建立一個混合式推薦模型,並比較不同推薦方法之間的成效。最後依預測推薦分數的高低來進行Top-N書籍推薦,找出最適宜推薦的前N名,提供使用者選擇時的參考。
    Developing new customers and retaining existing ones is a key issue in the development of online shopping platforms. The research is a domestic online shopping platform that sells books. Most of these products are purchased in small quantities and have special needs. For consumers who like to read, the number of purchases and the amount will be several times higher than most people. Finding consumers who like reading, regulatory activities or personalized product recommendations can not only increase the sales of goods, but also reduce marketing costs.
    In this study, RFM is used to analyze the consumption of existing customers, and Cascade K-means is used for customer grouping. Referring to the customer value matrix, the grouping results are divided into four groups to find the best customers, and finally the user-based and Project-based best predictive scores are used to calculate different weights, establish a hybrid recommendation model, and compare the effectiveness of different recommendation methods. Finally, the Top-N book recommendation is based on the predicted recommendation score, providing a reference for the user to select.
    显示于类别:[資訊管理學系暨資訊管理研究所 ] 博碩士論文

    文件中的档案:

    档案 描述 大小格式浏览次数
    index.html0KbHTML147检视/开启


    在CCUR中所有的数据项都受到原著作权保护.


    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - 回馈