In this study, we propose a structured methodology that makes finding experts on given topics much easier. The rapid growth of social networking websites has motivated us to find experts on such sites as Twitter, Facebook, and LinkedIn for crowdsourcing. Currently, Twitter is the largest microblog with an enormous amount of data available to be mined. There are more than 500 million registered users, and the number of active users is almost half of this. Every second, more than 9 thousand tweets are added to its data repository. For this reason, we apply our expert finding method to Twitter. We employ a set of tweets with their categories (i.e., Technology, Politics, Travel, Art, Music, Sports, Shopping, and Nature) to build our classifier based on LingPipe. Our goal is to find experts on each topic considering their tweets classified by our classifier. We also present a formula for ranking computation that takes into account the last five days’ classified tweets, all-time classified tweets, and activity of the users, and thoroughly discuss our method for determining the optimal coefficients in the presented formula for ensuring the quality of the selected experts.
Finding experts for better crowdsourcing
IPEK, OSMAN UYGUR
2012/2013
Abstract
In this study, we propose a structured methodology that makes finding experts on given topics much easier. The rapid growth of social networking websites has motivated us to find experts on such sites as Twitter, Facebook, and LinkedIn for crowdsourcing. Currently, Twitter is the largest microblog with an enormous amount of data available to be mined. There are more than 500 million registered users, and the number of active users is almost half of this. Every second, more than 9 thousand tweets are added to its data repository. For this reason, we apply our expert finding method to Twitter. We employ a set of tweets with their categories (i.e., Technology, Politics, Travel, Art, Music, Sports, Shopping, and Nature) to build our classifier based on LingPipe. Our goal is to find experts on each topic considering their tweets classified by our classifier. We also present a formula for ranking computation that takes into account the last five days’ classified tweets, all-time classified tweets, and activity of the users, and thoroughly discuss our method for determining the optimal coefficients in the presented formula for ensuring the quality of the selected experts.File | Dimensione | Formato | |
---|---|---|---|
Thesis.pdf
non accessibile
Descrizione: Revised version
Dimensione
2.08 MB
Formato
Adobe PDF
|
2.08 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/88338