Application of Refined K-means Clustering to the Implementation of Personalized Musical Recommendation System
碩士 === 國立高雄應用科技大學 === 電子工程系 === 100 === As digitized content and the Internet evolve in recent years, the number of users who obtain information via the Internet keeps growing. In the era of information explosion, the personal computer is no longer the only option as network terminals. A wide range...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
101
|
Online Access: | http://ndltd.ncl.edu.tw/handle/97793816946255882008 |
id |
ndltd-TW-100KUAS8393007 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-100KUAS83930072015-10-13T22:01:09Z http://ndltd.ncl.edu.tw/handle/97793816946255882008 Application of Refined K-means Clustering to the Implementation of Personalized Musical Recommendation System 應用改良式K-means分群法於個人化音樂推薦服務系統之實現 Zhen-Ming Wu 吳振銘 碩士 國立高雄應用科技大學 電子工程系 100 As digitized content and the Internet evolve in recent years, the number of users who obtain information via the Internet keeps growing. In the era of information explosion, the personal computer is no longer the only option as network terminals. A wide range of 3C products can communicate and link together with each other via the Internet. In this study, a music recommendation system based on Android operation system and mobile handhold devises is proposed. The objective is to establish a cloud computing service such that users can get the music matching information via music recommendation model. Two techniques are involved in our recommendation system: Collaborative Filter Model and Content-Base Filter Model. The rating information is collected by the users’ actual rating and the users with similar interest or preference will be found by system via collaborative filter model. An improved K-Means Clustering algorithm is proposed in this thesis which improves the similarity among data in the same group. The clustering classifies users into groups with the same preference. Recommendations are made base on the music preference of users in the same group. The candidates with highest score will be recommended to the users by the cloud recommendation system. The proposed improvement was experimented against previous approaches. The aggregation rate and discrimination rate were used as the performance indices. And then the Root Mean Square Error was used to evaluate the effectiveness of the recommendation system. Experiment results reveal that the proposed refinement outperforms previous schemes. Bin-Yih Liao Chin-Shiuh Shieh 廖斌毅 謝欽旭 101 學位論文 ; thesis 75 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立高雄應用科技大學 === 電子工程系 === 100 === As digitized content and the Internet evolve in recent years, the number of users who obtain information via the Internet keeps growing. In the era of information explosion, the personal computer is no longer the only option as network terminals. A wide range of 3C products can communicate and link together with each other via the Internet. In this study, a music recommendation system based on Android operation system and mobile handhold devises is proposed. The objective is to establish a cloud computing service such that users can get the music matching information via music recommendation model.
Two techniques are involved in our recommendation system: Collaborative Filter Model and Content-Base Filter Model. The rating information is collected by the users’ actual rating and the users with similar interest or preference will be found by system via collaborative filter model. An improved K-Means Clustering algorithm is proposed in this thesis which improves the similarity among data in the same group. The clustering classifies users into groups with the same preference. Recommendations are made base on the music preference of users in the same group. The candidates with highest score will be recommended to the users by the cloud recommendation system.
The proposed improvement was experimented against previous approaches. The aggregation rate and discrimination rate were used as the performance indices. And then the Root Mean Square Error was used to evaluate the effectiveness of the recommendation system. Experiment results reveal that the proposed refinement outperforms previous schemes.
|
author2 |
Bin-Yih Liao |
author_facet |
Bin-Yih Liao Zhen-Ming Wu 吳振銘 |
author |
Zhen-Ming Wu 吳振銘 |
spellingShingle |
Zhen-Ming Wu 吳振銘 Application of Refined K-means Clustering to the Implementation of Personalized Musical Recommendation System |
author_sort |
Zhen-Ming Wu |
title |
Application of Refined K-means Clustering to the Implementation of Personalized Musical Recommendation System |
title_short |
Application of Refined K-means Clustering to the Implementation of Personalized Musical Recommendation System |
title_full |
Application of Refined K-means Clustering to the Implementation of Personalized Musical Recommendation System |
title_fullStr |
Application of Refined K-means Clustering to the Implementation of Personalized Musical Recommendation System |
title_full_unstemmed |
Application of Refined K-means Clustering to the Implementation of Personalized Musical Recommendation System |
title_sort |
application of refined k-means clustering to the implementation of personalized musical recommendation system |
publishDate |
101 |
url |
http://ndltd.ncl.edu.tw/handle/97793816946255882008 |
work_keys_str_mv |
AT zhenmingwu applicationofrefinedkmeansclusteringtotheimplementationofpersonalizedmusicalrecommendationsystem AT wúzhènmíng applicationofrefinedkmeansclusteringtotheimplementationofpersonalizedmusicalrecommendationsystem AT zhenmingwu yīngyònggǎiliángshìkmeansfēnqúnfǎyúgèrénhuàyīnlètuījiànfúwùxìtǒngzhīshíxiàn AT wúzhènmíng yīngyònggǎiliángshìkmeansfēnqúnfǎyúgèrénhuàyīnlètuījiànfúwùxìtǒngzhīshíxiàn |
_version_ |
1718071240628895744 |