具個人化特色的智慧型點歌系統

Chaoyang University of Technology Institutional Repository > 資訊學院 > 資訊管理系、資訊科技研究所 > 博碩士論文 > Item 310901800/26440

Please use this identifier to cite or link to this item: http://ir.lib.cyut.edu.tw:8080/handle/310901800/26440

Title:	具個人化特色的智慧型點歌系統 The Intelligent Song Requesting System with Personalized Characteristics
Authors:	黃若嘉 Huang, Ruo-Jia
Contributors:	資訊管理系碩士班陳榮昌 Rong-Chung Chen
Keywords:	點歌;協同過濾;模糊推論;適性化推薦系統;知識庫 Information Retrieval;Requesting Song;Knowledge Base;Fuzzy Inference;Adaptation Recommendation System
Date:	2007-12-31
Issue Date:	2016-01-08 15:33:24 (UTC+8)
Abstract:	隨著電子商務的蓬勃發展，線上服務機制的功能愈來愈趨於多元化，使用者可以透過網際網路隨時隨地點歌(Requesting Song)歡唱。目前網路線上點歌系統大多利用新進歌曲點歌、或熱門點唱排行榜點歌等方式供使用者點選，在眾多的資訊中，使用者必須主動挖掘其所要的資訊，例如，透過歌曲編號、關鍵字搜尋等進行資訊的篩選。然而，上述等功能無法考慮到個別使用者的興趣差異，無法有效的提供使用者個人化的推薦。首先，我們提出利用協同過濾(Collaborative Filtering)推薦的方法，主動進行個人化之推薦，依據每個使用者過去的點選紀錄以及使用者輪廓中的興趣差異，搜尋出每個使用者的同好，以同好名單進行推薦(Recommendation)。並且結合模糊推論(Fuzzy Inference)所建置的知識庫(Knowledge Base)，將不同的特徵類別以模糊法則的方式呈現出來，每一個特徵類別會對應到一個模糊法則，因此，從過去的點歌紀錄中，當某一個使用者在某些特徵類別上有特殊傾向時，該模糊法則就會被觸發而形成推薦的依據。最後，我們透過隱性回饋的方式來加強知識庫媒合的精確度，除了可以使系統更有效率外，並可以從經驗中學習，以提供不同興趣傾向的使用者的適性化推薦系統(Adapted Recommendation System)。 Most of the requesting song system over the networks could only use the listing of new songs or the billboard of the hot requesting songs to speed up the requesting time. The users could only use the keywords or I.D. numbers of the songs to request their interested songs by their experiences. It will be more convenient for the users if these experiences can be accumulated in the requesting song system. Thus, we attempt to apply the method of information retrieval to find the trends of requesting song from their requesting records so as to provide the personalized recommendation. In our research, the fuzzy inference rules are combined to establish the knowledge base. The various attributive classes are represented as fuzzy inference rules. So, the fuzzy rule will be triggered while the user had the inclination of the corresponding attributive class in his previous requesting records. Finally, the recommendation is accepted by user or not will be feedback to the system to enhance the accuracy of the recommendation. Thus, we offer an adapted recommendation system for the variety of trends of the users.
Appears in Collections:	[資訊管理系、資訊科技研究所] 博碩士論文

Files in This Item:

File	Size	Format
095CYUT5396011-001.zip	710Kb	Unknown	70	View/Open

Loading...