Abstract
Cultural and scientific content is available in different repositories, distributed within archives, libraries and museum collections across Europe. These collections encompass different kinds of data, among them scientific publications, audio files, images and videos. While such cultural and scientific content is usually openly accessible, accessing it takes efforts and requires in-depth knowledge about the existing repositories and portals – a major hurdle that not all users can successfully overcome. Therefore, EEXCESS pursues the idea of giving users effortless access to Europe's rich digitally available, cultural heritage content. Through contextualized and personalized recommendations of cultural and scientific content we aim to bring the content to the user instead of bringing the users to the content.
In order to realize contextualized and personalized recommendations, EEXCESS needs to i) harmonise cultural and scientific metadata, ii) ensure easy integration of new repositories, iii) realize a unique access point via source federation and iv) inject the content in the users' digital whereabouts in a pleasant and personalised way. To this end, EEXCESS researched and implemented different components, all of them available as open source. The main components are:
Metadata mapping: The metadata mapping is responsible for harmonising different metadata source-formats into one unified EEXCESS metadata format, which is an adaption of the Europeana EDM format and geared towards recommendations.
Federated search: The federated search engine distributes the search request and the user profile to all connected partner systems, and aggregates the results based on the user profile. Currently EEXCESS provides federated access to over 140 millionen scientific and cultural objects aggregated from large aggregators like Europeana or Mendeley as well as from smaller memory organisations like for example the KIM Portal in Baselland.
User context detection and profiling: The user context detection component is a client component responsible for detecting the users information need, generating a suitable search query and inferring a descriptive user profile from the digital user behaviour.
Content injection and visualisation: The injection components are responsible for including the retrieved content in a suitable manner in the user's current digital whereabouts, e.g. into the blog entry, the user is currently editing. Visualisations support the investigation and filtering of retrieved results.
Privacy preservation: Because the user context detection collects information about user behaviour, EEXCESS implements a privacy-preserving, personalized search to ensure user privacy. Search requests and user profiles are sent through a privacy proxy component ensuring unlinkability (detaching the query from the actual user) and indistinguishability (hiding a single user within a larger group) of requests.
EEXCESS has developed different clients suitable for different user tasks. Content creation gives access to cultural heritage content while writing and is available as Google Docs plugin, a Moodle server plugin and a Wordpress Plugin. Content consumption automatically enriches web pages with cultural and scientific content when users browse the Web through a Chrome Browser extension.
In order to realize contextualized and personalized recommendations, EEXCESS needs to i) harmonise cultural and scientific metadata, ii) ensure easy integration of new repositories, iii) realize a unique access point via source federation and iv) inject the content in the users' digital whereabouts in a pleasant and personalised way. To this end, EEXCESS researched and implemented different components, all of them available as open source. The main components are:
Metadata mapping: The metadata mapping is responsible for harmonising different metadata source-formats into one unified EEXCESS metadata format, which is an adaption of the Europeana EDM format and geared towards recommendations.
Federated search: The federated search engine distributes the search request and the user profile to all connected partner systems, and aggregates the results based on the user profile. Currently EEXCESS provides federated access to over 140 millionen scientific and cultural objects aggregated from large aggregators like Europeana or Mendeley as well as from smaller memory organisations like for example the KIM Portal in Baselland.
User context detection and profiling: The user context detection component is a client component responsible for detecting the users information need, generating a suitable search query and inferring a descriptive user profile from the digital user behaviour.
Content injection and visualisation: The injection components are responsible for including the retrieved content in a suitable manner in the user's current digital whereabouts, e.g. into the blog entry, the user is currently editing. Visualisations support the investigation and filtering of retrieved results.
Privacy preservation: Because the user context detection collects information about user behaviour, EEXCESS implements a privacy-preserving, personalized search to ensure user privacy. Search requests and user profiles are sent through a privacy proxy component ensuring unlinkability (detaching the query from the actual user) and indistinguishability (hiding a single user within a larger group) of requests.
EEXCESS has developed different clients suitable for different user tasks. Content creation gives access to cultural heritage content while writing and is available as Google Docs plugin, a Moodle server plugin and a Wordpress Plugin. Content consumption automatically enriches web pages with cultural and scientific content when users browse the Web through a Chrome Browser extension.
Original language | English |
---|---|
Journal | D-Lib Magazine |
Volume | 22 |
Issue number | 3/4 |
DOIs | |
Publication status | Published - 1 Mar 2016 |
Externally published | Yes |