Zur Seitenansicht
 

Titelaufnahme

Titel
Leveraging Topic Extraction for Hashtag Recommendations
VerfasserSchneider, Georg
Betreuer / BetreuerinZangerle, Eva
Erschienen2015
HochschulschriftInnsbruck, Univ., Masterarb., 2015
Anmerkung
Arbeit an der Bibliothek noch nicht eingelangt - Daten nicht geprüft
Datum der AbgabeJuli 2015
SpracheEnglisch
DokumenttypMasterarbeit
Schlagwörter (EN)hashtag / recommender / recommender system / recommendation / hashtag recommendation / hashtag recommender / hybrid recommender / twitter / Maui / topic extraction / keyword extraction
URNurn:nbn:at:at-ubi:1-4828 Persistent Identifier (URN)
Lizenz
CC-BY-NC-Lizenz (4.0)Creative Commons Namensnennung - Nicht kommerziell 4.0 International Lizenz
Zugriffsbeschränkung
 Das Werk ist frei verfügbar
Dateien
Leveraging Topic Extraction for Hashtag Recommendations [0.78 mb]
Links
Nachweis
Klassifikation
Zusammenfassung (Deutsch)

The microblogging service Twitter uses hashtags to tag and categorize messages. To automatically recommend appropriate hashtags, we analyze the text contained in websites linked in a Twitter message. This paper compares and analyzes various approaches and algorithms for topic extraction and seeks to optimize and use them in order to generate keywords for a Twitter message. These keywords, along with the text of the input message and any hashtags in it, are then used by six recommender algorithms and three hybrid recommenders to find and suggest suitable hashtags.

Zusammenfassung (Englisch)

The microblogging service Twitter uses hashtags to tag and categorize messages. To automatically recommend appropriate hashtags, we analyze the text contained in websites linked in a Twitter message. This paper compares and analyzes various approaches and algorithms for topic extraction and seeks to optimize and use them in order to generate keywords for a Twitter message. These keywords, along with the text of the input message and any hashtags in it, are then used by six recommender algorithms and three hybrid recommenders to find and suggest suitable hashtags.