A Framework for Time-Aware Recommendations

Stefanidis, Kostas; Ntoutsi, Irene; Nørvåg, Kjetil; Kriegel, Hans‐Peter

doi:10.1007/978-3-642-32597-7_29

Cited by 7 publications

(6 citation statements)

References 18 publications

(22 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Stefanidis et al [24] presented a framework for generating recommendations based on users' recent preferences and providing different suggestions under different temporal specifications. Gao et al [25] presented a model for location-based recommendations to a user based on his personal preferences, to facilitate his exploration of new areas of a city.…”

Section: A Temporal Collaborative Filtering Methodsmentioning

confidence: 99%

“…In our experiments, we used a time window equal to a semiannual, where as training set we considered all the past months of the previous semiannuals and the first five months of the current ongoing semiannual. Therefore, we have nine different test sets of tuples at test months 6,12,18,24,30,36,42,48, and 54, denoted by red lines in Fig. 3, and nine different training sets of tuples at the respective past months.…”

Section: A Lastfm Datasetmentioning

confidence: 99%

See 1 more Smart Citation

Modeling Users Preference Dynamics and Side Information in Recommender Systems

Rafailidis

Νανόπουλος

2016

IEEE Trans. Syst. Man Cybern, Syst.

View full text Add to dashboard Cite

In recommender systems user preferences can be fairly dynamic, as users tend to exploit a wide range of items and modify their tastes accordingly over time. In this paper, we model user-item interactions over time using a tensor that has time as a dimension (mode). To account for the fact that user preferences change individually, we propose a new measure of user-preference dynamics (UPD) that captures the rate with which the current preferences of each user have been shifted. UPD shows the variability in how users interact with items in recommender systems. We generate recommendations based on a tensor factorization technique, where the importance of past user preferences are weighted according to their UPD values, that is, higher UPD values downweigh more past user preferences. Additionally, we exploit users' side data, such as demographics, which improve the accuracy of recommendations based on a coupled tensormatrix factorization scheme. Our empirical evaluation uses two real benchmark datasets from the social media platforms Last.fm and MovieLens, containing users' history records pertaining to listening to songs and viewing movies, respectively. We demonstrate that in both datasets, there are users with a varying level of dynamics, expressed by the UPD metric. Our experimental results show that the proposed method outperforms several baselines, by taking into account both dynamics and side data of users.Index Terms-Coupled tensor factorization (CTF), recommender systems, users' dynamics. , Cyprus. His current research interests include databases, multimedia information retrieval, social media mining, and parallel and cloud computing.

show abstract

Section: A Temporal Collaborative Filtering Methodsmentioning

confidence: 99%

Section: A Lastfm Datasetmentioning

confidence: 99%

Modeling Users Preference Dynamics and Side Information in Recommender Systems

Rafailidis

Νανόπουλος

2016

IEEE Trans. Syst. Man Cybern, Syst.

View full text Add to dashboard Cite

show abstract

“…With respect to the management of time dimension of data, several techniques are based on either discrete time windows, where only data in sliding windows are considered (eg, the latest n ratings; all rating in the last t hours); or continuous decay functions, that give more weight to recently rated items by a user . Considering our proposal, we define an approach that adopt, as also done in a previous study, the two previous different types in the recommendation process. The combination of two techniques ensures us to get always up‐to‐date data by using the sliding window, and makes sure that the recent posts are more relevant than old ones, by means of the decay function.…”

Section: Related Workmentioning

confidence: 99%

“…The aggregation follows an aging operation which weights the topic interests of users based on newness (freshness) of a post. In particular, as in a previous study, we adopt 2 different types of aging mechanisms into the recommendation process: a sliding window model and a damped window model (ie, a decay function). The sliding window should ensure having always up‐to‐date data, and the decay function makes sure that the recent posts are more relevant than old ones.…”

Section: Overall Approachmentioning

confidence: 99%

Social media marketing through time‐aware collaborative filtering

Maio

Fenza

Gallo

et al. 2017

Concurrency and Computation

View full text Add to dashboard Cite

SummarySocial media is assuming a crucial role in purchasing decisions, and most companies are using social media for marketing. Making sense of the unstructured information content shared by users along the time is an emerging challenge. We advise that the dynamic nature with which trends and user's interests evolve along the timeline requires the revising of well-assessed methods to address, for instance, recommendation provisioning, information retrieval, and so on. ie, Twitter. 6 In particular, we advise that time awareness is crucial to more effectively estimate user's interests, and thus, to better address social media marketing improving "vanity metrics" of likes and shares that in social marketing remain the most commonly used measures for evaluating success on social media attempting to increase traction, for instance, posting the right message at the right time.Problem. Let us consider a timestamped tweet stream = {tw 1 , … , tw n } in the time window starting at time t s and ending at time t e , and let us suppose topics corresponding to the content of each

show abstract

“…Typically, recommendation approaches are distinguished between: content-based, that recommend items similar to those the user previously preferred [17], collaborative filtering, that recommend items that users with similar preferences liked [14,7] and hybrid ones [4]. Several extensions have been proposed, such as employing multi-criteria ratings [2] or further contextual information [20], and providing time-aware recommendations [26,23]. Recently, there are also approaches focusing on extending database queries with recommendations [15,22].…”

Section: Related Workmentioning

confidence: 99%

Fast Group Recommendations by Applying User Clustering

Ntoutsi

Stefanidis

Nørvåg

et al. 2012

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Abstract. Recommendation systems have received significant attention, with most of the proposed methods focusing on personal recommendations. However, there are contexts in which the items to be suggested are not intended for a single user but for a group of people. For example, assume a group of friends or a family that is planning to watch a movie or visit a restaurant. In this paper, we propose an extensive model for group recommendations that exploits recommendations for items that similar users to the group members liked in the past. We do not exhaustively search for similar users in the whole user base, but we pre-partition users into clusters of similar ones and use the cluster members for recommendations. We efficiently aggregate the single user recommendations into group recommendations by leveraging the power of a top-k algorithm. We evaluate our approach in a real dataset of movie ratings.

show abstract

A Framework for Time-Aware Recommendations

Cited by 7 publications

References 18 publications

Modeling Users Preference Dynamics and Side Information in Recommender Systems

Modeling Users Preference Dynamics and Side Information in Recommender Systems

Social media marketing through time‐aware collaborative filtering

Fast Group Recommendations by Applying User Clustering

Contact Info

Product

Resources

About