Releases
1.0
EuropeanaRS v1.0
Hybrid Recommender System based on:
Learning Object Profile Similarity
User Profile Similarity
Quality
Popularity
Implementations to calculate similarities:
Text semantic similarities calculated using cosine similarity distance based on the TF-IDF .
Numeric similarities calculated as arithmetic distance in a specific scale.
Categorial fields (and booleans) calculated with equality functions.
Filtering recommendations based on Learning Object similarity, User Profile similarity, quality or popularity.
Filtering recommendations based o specific fields. For instance, filter Learning Objects when title similarity is less than 0.5.
Customizable weights:
General (Learning Object similarity, User Profile similarity, quality or popularity).
Field specific.
Customizable thresholds for filters:
General (Learning Object similarity, User Profile similarity, quality or popularity).
Field specific.
Search Engine based on sphinx.
Management of Learning Objects, Learning Object Profiles, Users, User Profiles and Applications.
EuropeanaRS API for delivering recommendations to third-party web client applications.
JavaScript library for web applications that want to use the EuropeanaRS API.
Fully customizable settings for the system.
At application level (system default settings)
At user level (user settings)
At EuropeanaRS API level (third-party web app settings)
Basic UI
Login with:
Registration in EuropeanaRS
Facebook (OAuth2)
Europeana (OAuth2)
Europeana (UserAuthentication API)
Internacionalization support. Supported languages: English, Spanish.
Europeana implementations for:
Use Europeana Search API
Use MyEuropeana API including the Europeana UserAuthentication service
OAI-PMH Europeana service (in beta)
A Europeana Mimic component to be used for developing purposes
Demos for using the EuropeanaRS API.
Dump with a dataset of more than 10.000 Learning Objects retrieved from Europeana for developing purposes.
You can’t perform that action at this time.