Temporal diversity in recommender systems

  title={Temporal diversity in recommender systems},
  author={Neal Lathia and Stephen Hailes and Licia Capra and X. Amatriain},
  journal={Proceedings of the 33rd international ACM SIGIR conference on Research and development in information retrieval},
  • N. Lathia, S. Hailes, X. Amatriain
  • Published 19 July 2010
  • Computer Science
  • Proceedings of the 33rd international ACM SIGIR conference on Research and development in information retrieval
Collaborative Filtering (CF) algorithms, used to build web-based recommender systems, are often evaluated in terms of how accurately they predict user ratings. However, current evaluation techniques disregard the fact that users continue to rate items over time: the temporal characteristics of the system's top-N recommendations are not investigated. In particular, there is no means of measuring the extent that the same items are being recommended to users over and over again. In this work, we… 

Figures from this paper

Exploiting the diversity of user preferences for recommendation
This paper proposes an adaptation of search result diversification methods to recommender systems based on query reformulation, which identifies the diversity within user profiles and generates partial recommendations based on homogeneous subsets of user preferences (sub-profiles), which are combined later to produce a final recommendation.
Evaluating collaborative filtering over time
Investigating collaborative filtering from a temporal perspective is not only more suitable to the context in which recommender systems are deployed, but also opens a number of future research opportunities.
Rating-Based Collaborative Filtering: Algorithms and Evaluation
The concepts, algorithms, and means of evaluation that are at the core of collaborative filtering research and practice are reviewed, and two more recent directions in recommendation algorithms are presented: learning-to-rank and ensemble recommendation algorithms.
Time-Aware Novelty Metrics for Recommender Systems
This paper proposes a formulation to capture the time-aware novelty (or freshness) of the recommendation lists, according to different time models of the items, and provides a measure to account for how much a system is promoting fresh items in its recommendations.
Goal-driven collaborative filtering
This thesis intends to study recommender systems from a goal-oriented point of view, where the recommendation goals are defined, their associated measures are built and a unified error minimisation framework is proposed that flexibly covers various (directional) risk preferences.
Diversity in Recommender Systems: Bridging the gap between users and systems
It is shown that combining existing recommendation selection methods is a good mean to obtain diversity in recommendation lists, and two sets of approaches from the literature are compared to underline that their results are significantly different.
Diversity in Recommender System
Objective functions that capture the trade-off between the goals and optimization problems associated with the maximization of these objectives are presented.
Controlling Consistency in Top-N Recommender Systems
This work analyzes how small changes in either the user profile or the recommender model may affect the consistency of Top-N recommendation systems and designs two mechanisms able to promote consistency without degrading accuracy and novelty of recommendations.
Analysing exposure diversity in collaborative recommender systems—Entropy fusion approach
Beyond Collaborative Filtering: The List Recommendation Problem
This work introduces the list recommendation problem and proposes a novel two-layered framework that builds upon existing CF algorithms to optimize a list's click probability and evaluates the approach using a novel adaptation of Inverse Propensity Scoring which facilitates off-policy estimation of the method's CTR and showcases its effectiveness in real-world settings.


Evaluating collaborative filtering recommender systems
The key decisions in evaluating collaborative filtering recommender systems are reviewed: the user tasks being evaluated, the types of analysis and datasets being used, the ways in which prediction quality is measured, the evaluation of prediction attributes other than quality, and the user-based evaluation of the system as a whole.
Hybrid Recommender Systems: Survey and Experiments
  • R. Burke
  • Computer Science
    User Modeling and User-Adapted Interaction
  • 2004
This paper surveys the landscape of actual and possible hybrid recommenders, and introduces a novel hybrid, EntreeC, a system that combines knowledge-based recommendation and collaborative filtering to recommend restaurants, and shows that semantic ratings obtained from the knowledge- based part of the system enhance the effectiveness of collaborative filtering.
Temporal collaborative filtering with adaptive neighbourhoods
This work formalises CF as a time-dependent, iterative prediction problem, and addresses the problem of parameter selection and update, and proposes a method to automatically assign and update per-user neighbourhood sizes that outperforms setting global parameters.
Confidence Displays and Training in Recommender Systems
Providing training about a confidence display to experienced users greatly reduced user satisfaction in the recommender system, and subtle effects about how and when to train users when adding features to a system are demonstrated.
Avoiding monotony: improving the diversity of recommendation lists
This work model the competing goals of maximizing the diversity of the retrieved list while maintaining adequate similarity to the user query as a binary optimization problem, leading to a parameterized eigenvalue problem whose solution is finally quantized to the required binary solution.
Collaborative filtering with temporal dynamics
Two leading collaborative filtering recommendation approaches are revamp and a more sensitive approach is required, which can make better distinctions between transient effects and long term patterns.
Toward the next generation of recommender systems: a survey of the state-of-the-art and possible extensions
This paper presents an overview of the field of recommender systems and describes the current generation of recommendation methods that are usually classified into the following three main
Similarity vs. Diversity
This paper proposes and evaluates strategies for improving retrieval diversity in CBR systems without compromising similarity or efficiency and argues that often diversity can be as important as similarity.
Diversifying search results
This work proposes an algorithm that well approximates this objective in general, and is provably optimal for a natural special case, and generalizes several classical IR metrics, including NDCG, MRR, and MAP, to explicitly account for the value of diversification.
Improving personalized web search using result diversification
Three methods to increase the diversity of the top results are proposed and evaluated and the effectiveness of these methods is evaluated.