A Multimedia User Preference Model that Supports Semantics and its Application to MPEG 7/21

Tsinaraki, Chrisa; Christodoulakis, Stavros

doi:10.1109/mmmc.2006.1651299

Cited by 22 publications

(29 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We should note that this evaluation is considerably different from conducting user studies to judge the performance of recommender systems and from modeling detailed user preferences recorded with ontology-based approaches [8,25,48]. The former, aside from the biases, is time-consuming and challenging.…”

Section: Ground Truthmentioning

confidence: 99%

Combining content with user preferences for non-fiction multimedia recommendation: a study on TED lectures

Pappas

Popescu-Belis

2014

Multimed Tools Appl

View full text Add to dashboard Cite

This paper introduces a new dataset and compares several methods for the recommendation of non-fiction audio visual material, namely lectures from the TED website. The TED dataset contains 1,149 talks and 69,023 profiles of users, who have made more than 100,000 ratings and 200,000 comments. The corresponding metadata, which we make available, can be used for training and testing generic or personalized recommender systems. We define content-based, collaborative, and combined recommendation methods for TED lectures and use cross-validation to select the best parameters of keyword-based (TFIDF) and semantic vector space-based methods (LSI, LDA, RP, and ESA). We compare these methods on a personalized recommendation task in two settings, a cold-start and a non-cold-start one. In the cold-start setting, semantic vector spaces perform better than keywords. In the noncold-start setting, where collaborative information can be exploited, content-based methods are outperformed by collaborative filtering ones, but the proposed combined method shows acceptable performances, and can be used in both settings. For the generic recommendation task, LSI and RP again outperform TF-IDF.Keywords Content-based multimedia indexing · Recommender systems · Multimedia recommendation · TED lectures We are grateful for the support received for this work from the European Union through the inEvent project FP7-ICT n. 287872 (see http://www.inevent-project.eu).

show abstract

Section: Ground Truthmentioning

confidence: 99%

Combining content with user preferences for non-fiction multimedia recommendation: a study on TED lectures

Pappas

Popescu-Belis

2014

Multimed Tools Appl

View full text Add to dashboard Cite

show abstract

“…Since this transformation is Schema dependent it is not automated within the Framework software. The reference ontology we used is an application of the DS-MIRF ontological infrastructure (Tsinaraki 2006) and the WordNet for the syntactic analysis. The repository for accessing the instances is the DS-MIRF Metadata Repository (Tsinaraki 2006).…”

Section: An Application Of the Ontonl Frameworkmentioning

confidence: 99%

“…The reference ontology we used is an application of the DS-MIRF ontological infrastructure (Tsinaraki 2006) and the WordNet for the syntactic analysis. The repository for accessing the instances is the DS-MIRF Metadata Repository (Tsinaraki 2006). The DS-MIRF OntoNL Manager provides the OntoNL component with the ontologies for the disambiguation and the natural language expression for disambiguation.…”

Section: An Application Of the Ontonl Frameworkmentioning

confidence: 99%

The OntoNL Semantic Relatedness Measure for OWL ontologies

Karanastasi

Christodoulakis

2007

2007 2nd International Conference on Digital Information Management

View full text Add to dashboard Cite

“…They extend the UserPreferences DS so that 'events' may be accommodated within the ClassificationPreferences DS and suitable programming recommended according to suitable matching events in the content metadata. More substantially, Tsinaraki and Christodoulakis [19,20] define a subtype of the MPEG-7 MDS CreationPreferencesType, the SCreationPreferencesType, in which they introduce a set of content preference description elements of type WeighedSemanticBagType, which is an extension of the MPEG-7 SemanticBagType but includes a preferenceValue attribute so that the relevant importance of a content preference description can be specified as an integer value in the range [−100, 100] (10 is the default value). Negative preferenceValues are used to denote fuzzy negation.…”

Section: Tsinaraki and Christodoulakismentioning

confidence: 99%

Closing the content-user gap in MPEG-7: the hanging basket model

Agius

Angelides

2007

Multimedia Systems

View full text Add to dashboard Cite

MPEG-7 is the leading standard for multimedia metadata creation, providing a vast array of description tools. Nevertheless, there exists a disparity between the content models and user models that may be created by the standard: metadata for content models is wide ranging and structured, but metadata for user models only support the specification of an extremely limited range of preferences where content metadata cannot be fully exploited. This results in a very incomplete mapping of user models to content models which we term the MPEG-7 content-user gap. We present a hanging basket model that closes the content-user gap by considering user models to be isomorphic to content models and thus enabling both models to be represented through MPEG-7 content description tools, specifically those concerned with events, objects and properties.

show abstract

A Multimedia User Preference Model that Supports Semantics and its Application to MPEG 7/21

Cited by 22 publications

References 12 publications

Combining content with user preferences for non-fiction multimedia recommendation: a study on TED lectures

Combining content with user preferences for non-fiction multimedia recommendation: a study on TED lectures

The OntoNL Semantic Relatedness Measure for OWL ontologies

Closing the content-user gap in MPEG-7: the hanging basket model

Contact Info

Product

Resources

About