Demand-Driven Tag Recommendation

Assessing the quality of textual features in social media

“…In other words, the idea is to expand the contents of the TAGS feature by introducing new recommended terms extracted from the other three features, and then evaluate the quality of such recommendations. We note that, despite the great variety of different tag recommendation methods and tag analyses available in the literature (Menezes et al, 2010;Sigurbjornsson & van Zwol, 2008;Belém et al, 2011;Rendle & Schmidt-Thie, 2010;Lipczak et al, 2009;Zhang et al, 2009), we are not aware of any previous effort in quantifying the quality of different sources of candidate terms for recommendation purposes. This is our objective here, which distinguishes the present effort from previous studies.…”

“…Sigurbjornsson and van Zwol (2008), for instance, exploited metrics of tag co-occurrence, applying them to an initial set of tags associated with the target object to produce a final ranking of candidate tags. Tag co-occurrence patterns are also exploited by Menezes et al (2010), although the authors use a lazy associative tag recommendation method in order to efficiently uncover more sophisticated patterns, which ultimately leads to superior results in comparison with the best method in (Sigurbjornsson & van Zwol, 2008). In Belém et al (2011) we extended traditional tag co-occurrence based methods to include not only tags that had been previously assigned to the objects but also terms extracted from other textual features, applying several heuristic metrics to capture the relevance of each candidate term as a recommendation for the target object.…”

“…The goal of this task is to support users during the process of tagging Web 2.0 objects, by recommending ''good'' (i.e., high quality) tags for describing an object (Guan et al, 2009;Song et al, 2008;Sigurbjornsson & van Zwol, 2008;Byde et al, 2007;Lipczak et al, 2009;Rendle & Schmidt-Thie, 2010;Belém et al, 2011;Menezes et al, 2010;Clements et al, 2010). We here focus on recommending tags because: (1) it is the most studied textual feature on the Web 2.0, being of more interest to the community; (2) our characterization and classification results indicate that this is a feature of quality (with respect to several aspects) when used in isolation; (3) recommending new terms to other features does not make much sense as they are typically composed of full sentences with semantic meanings, while the TAGS feature contains set of terms; and (4) the recommended tags can be, at least partially, automatically evaluated, as we shall see.…”

Exploiting relevance, novelty and diversity in tag recommendation

“…As métricas de relevância aqui exploradas foram previamente aplicadas na tarefa de recomendação de tags [3,12,11,8]. Não as apresentamos aqui por questões de espaço.…”

Improving daily deals recommendation using explore-then-exploit strategies

“…Finally, customers' taste and interest may undergo temporal drifts as new deals appear (Byers et al 2012a;Lappas and Terzi 2012). Previous results demonstrate the effectiveness of recommendation algorithms in application scenarios as diverse as movie (Miller et al 2003) and music recommendation (Aizenberg et al 2012;Koenigstein et al 2011), scientific paper assignment (Conry et al 2009), and tag recommendation (Menezes et al 2010), but leave open the question of whether existing recommendation algorithms are effective for the challenging scenario posed by daily-deals recommendation. More specifically, it is still unclear whether these algorithms are able to improve over the naive strategy based on sending non-personalized email messages with ordinary deals.…”