應用資料探勘技術於社群網路廣告回應之分析與預測

CCUR > College of Business > Department of Information Management & Graduate Institute of Information Management > Thesis > Item 987654321/30232

Please use this identifier to cite or link to this item: https://irlib.pccu.edu.tw/handle/987654321/30232

Title:	應用資料探勘技術於社群網路廣告回應之分析與預測 Applying Data Mining Technology for Analysis and Prediction of Advertisement Response in Social Networks
Authors:	林浩緯 Lin, Hao Wei
Contributors:	資訊管理學系
Keywords:	決策樹邏輯斯迴歸社群網路廣告行銷 decision tree logistic regression social network advertising and marketing
Date:	2015-06
Issue Date:	2015-08-17 15:38:05 (UTC+8)
Abstract:	隨著網際網路的普及化，對於現代的人們幾乎離不開網路世界，網路已成為日常生活中不可或缺的一部分。也越來越多社群網路資訊的交流與分享提供了新的途徑。透過社群網路平台提供使用者瞭解產品功能，並使消費者得到最大的滿足，找到消費者感興趣或是相關性高的資訊。因此自從社群網路問世以來，許多企業主與廣告商一直在研究瞭解消費者對於感興趣產品與喜好需求。本研究利用決策樹C4.5與CART方式找出對於使用者在社群網路關鍵屬性特徵，經由C4.5與CART十折交叉驗證方式比較是否有相同屬性特徵，接著透過屬性特徵找出使用者廣告回應接受度高的使用因素，最後使用決策樹J48(decision tree)與邏輯斯迴歸(logistic regression)兩種技術方法建立預測個人化廣告回應模型，實驗結果顯示以邏輯斯迴歸預測優於使用決策樹的方式。 As the popularity of Internet access grows, Internet has become an essential in their daily life for people who rely on it. More and more community network provides a new way to exchanging and sharing information. The community network platform provides product information and functions that make consumer to get the greatest satisfaction, or to find out consumers interests and highly related products. Since the advent of the Internet community, many business owners and advertisers have been researching to understand consumer needs and preferences for products of interest. In this research, we use Decision tree C4.5 and CART for consumers to identify the key attributes of social media features that highly affect the response rate of users to web advertising. Finally, the advertising forecasting models are build for each user using two techniques: Decision tree and Logistic regression. Experimental results show that Logistic regression outperforms Decision tree.
Appears in Collections:	[Department of Information Management & Graduate Institute of Information Management] Thesis

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	426	View/Open

View Licence

Loading...