東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

On Music and Social Event Recommende...

Ding, Hao.

Linked to FindBook

Google Book

Amazon

博客來

On Music and Social Event Recommender Systems.

Record Type:	Electronic resources : Monograph/item
Title/Author:	On Music and Social Event Recommender Systems./
Author:	Ding, Hao.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 2017,
Description:	129 p.
Notes:	Source: Dissertations Abstracts International, Volume: 79-01, Section: B.
Contained By:	Dissertations Abstracts International79-01B.
Subject:	Social research. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10281391
ISBN:	9781369850505

On Music and Social Event Recommender Systems.
Ding, Hao.

On Music and Social Event Recommender Systems. - Ann Arbor : ProQuest Dissertations & Theses, 2017 - 129 p.

Source: Dissertations Abstracts International, Volume: 79-01, Section: B.

Thesis (Ph.D.)--Polytechnic Institute of New York University, 2017.

This item must not be sold to any third party vendors.

Collaborative filtering (CF) is the state-of-art technique of recommender system. However, it can neither recommend new items with no user feedbacks, nor can it recommend "long-tail" unpopular items with good accuracy. Content-based filtering can solve both problems through content analysis. However, content-based filtering alone has a much worse performance than CF. To avoid defects from both approaches in the context of music recommendation, we fuse user feedbacks and content analysis into the probabilistic matrix factorization framework. In particular, we propose a recursive dynamic programming approach to computing item similarity matrix from item content. Item latent factors are predicted from the item similarity matrix when no usage data is available. We investigate how performance of recommendation algorithms vary on items with different popularities. Results show that our approach has better performance than the same hybrid model with simple item similarity measures and Matrix Factorization. Most of research on music recommendation worked on offline datasets. However, the rating values in an offline datasets have been influenced by the recommender system embedded on the website where the dataset came from. To avoid these defects caused by offline datasets, we designed an online 3rd-party music video application on Facebook website, where different recommenders can be implemented and compared using A/B testing method. Aside from the traditional online social networks, event-based Social Networks (EBSN) have experienced rapid growth in recent years and are changing people's ways of social interactions. User recommendation in EBSN is to recommend a list of users who are most likely to attend a new event. Due to the nature of new event and severe data sparsity, the traditional recommender systems do not work well for user recommendation. In the context of Meetup social platform, we first conduct a study of Meetup users to understand the major factors impacting their event participation decisions. We then develop a sliding-window based machine-learning model that effectively combines user features from multiple channels to recommend users to new events. Through evaluation using the Meetup dataset, we demonstrate that our model can capture the short-term consistency of user preferences and outperforms the traditional popularity-based and nearest-neighbor based recommendation models. Our model is suitable for real-time recommendation on practical EBSN platforms.

ISBN: 9781369850505Subjects--Topical Terms:

2122687
Social research.
Subjects--Index Terms:

Event recommendation

On Music and Social Event Recommender Systems.
LDR:03704nmm a2200373 4500 001 2270826
005 20201007134022.5
008 220629s2017 ||||||||||||||||| ||eng d
020 $a 9781369850505
035 $a (MiAaPQ)AAI10281391
035 $a (MiAaPQ)poly:10357
035 $a AAI10281391
040 $a MiAaPQ $c MiAaPQ
100 1 $a Ding, Hao. $3 2095867
245 1 0 $a On Music and Social Event Recommender Systems.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2017
300 $a 129 p.
500 $a Source: Dissertations Abstracts International, Volume: 79-01, Section: B.
500 $a Publisher info.: Dissertation/Thesis.
500 $a Advisor: Liu, Yong.
502 $a Thesis (Ph.D.)--Polytechnic Institute of New York University, 2017.
506 $a This item must not be sold to any third party vendors.
520 $a Collaborative filtering (CF) is the state-of-art technique of recommender system. However, it can neither recommend new items with no user feedbacks, nor can it recommend "long-tail" unpopular items with good accuracy. Content-based filtering can solve both problems through content analysis. However, content-based filtering alone has a much worse performance than CF. To avoid defects from both approaches in the context of music recommendation, we fuse user feedbacks and content analysis into the probabilistic matrix factorization framework. In particular, we propose a recursive dynamic programming approach to computing item similarity matrix from item content. Item latent factors are predicted from the item similarity matrix when no usage data is available. We investigate how performance of recommendation algorithms vary on items with different popularities. Results show that our approach has better performance than the same hybrid model with simple item similarity measures and Matrix Factorization. Most of research on music recommendation worked on offline datasets. However, the rating values in an offline datasets have been influenced by the recommender system embedded on the website where the dataset came from. To avoid these defects caused by offline datasets, we designed an online 3rd-party music video application on Facebook website, where different recommenders can be implemented and compared using A/B testing method. Aside from the traditional online social networks, event-based Social Networks (EBSN) have experienced rapid growth in recent years and are changing people's ways of social interactions. User recommendation in EBSN is to recommend a list of users who are most likely to attend a new event. Due to the nature of new event and severe data sparsity, the traditional recommender systems do not work well for user recommendation. In the context of Meetup social platform, we first conduct a study of Meetup users to understand the major factors impacting their event participation decisions. We then develop a sliding-window based machine-learning model that effectively combines user features from multiple channels to recommend users to new events. Through evaluation using the Meetup dataset, we demonstrate that our model can capture the short-term consistency of user preferences and outperforms the traditional popularity-based and nearest-neighbor based recommendation models. Our model is suitable for real-time recommendation on practical EBSN platforms.
590 $a School code: 1540.
650 4 $a Social research. $3 2122687
650 4 $a Computer science. $3 523869
653 $a Event recommendation
653 $a Music recommendation
653 $a Recommender systems
653 $a Social networks
690 $a 0344
690 $a 0984
710 2 $a Polytechnic Institute of New York University. $b Electrical and Computer Engineering. $3 1669328
773 0 $t Dissertations Abstracts International $g 79-01B.
790 $a 1540
791 $a Ph.D.
792 $a 2017
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10281391