HermesThe Hermes news personalization software is a Java-based tool that makes use of various Semantic Web technologies for news recommendation. Similar to most recommender systems, it operates based on user profiles. News messages are retrieved from RSS feeds, and processing steps make use of an OWL domain ontology that is constructed by domain experts. News items are classified using the GATE natural language processing software and the WordNet semantic lexicon.

The software has served as a basis for many projects, e.g., result ranking algorithms, graphical query languages, information extraction languages, news recommendation algorithms, etc. On this page, two major releases are available for non-commercial and private use.

Hermes 1.1 - HGQL

This software, programmed in Java by Arnout Verheij and Allard Kleijn, builds on the Hermes news personalization framework. It implements the Hermes Graphical Query Language (HGQL). HGQL enables users to create structured queries that use disjunctive, conjunctive, negation, and pattern operators. Also, the software implements various result ranking algorithms. Our software can be downloaded here as a zip-file, which can be extracted to any desired location on your harddrive.

Hermes 2.3 - Rule Learning

This software, programmed in Java by Wouter IJntema, builds on the Hermes news personalization framework. It implements the lexico-semantic Hermes Information Extraction Language (HIEL) for extracting events from news, and allows for automatic rule learning through a genetic programming approach. Our software can be downloaded here as a zip-file, which can be extracted to any desired location on your harddrive.

Software

To run Hermes, execute the following command line in the Hermes root folder:

java -jar Hermes.jar

Non-commercial and private use of the software is allowed. When publishing work that is based on my software, please ensure that proper citations are included. In case of questions, comments, or suggestions, feel free to contact me. Also, please notify me when you have developed extensions, additional components, etc., so that I can add them to the next software release.

Contact

Frederik Hogenboom, Ph.D.

Affiliation:
Econometric Institute
Erasmus School of Economics
Erasmus University Rotterdam

Postal:
P.O. Box 1738
NL-3000 DR Rotterdam
The Netherlands

Visiting:
Burgemeester Oudlaan 50
NL-3062 PA Rotterdam
The Netherlands

Room: H11-02 (map)
Phone: +31 (0)10 408 1264
Fax: +31 (0)10 408 9162
E-mail: fhogenboom@ese.eur.nl