Projects


SWEPS (Semantic Web Enhanced Product Search)

This projects aims to investigate the use of Semantic Web vocabularies (e.g., Google Rich Snippets) for efficient product discovery and presentation. For this, intelligent methods are developed that retrieve, integrate, aggregate, and present product information on the Web in order to reduce the consumer search effort for relevant products. The SWEPS project investigates how product annotation on the Web can facilitate retrieval, aggregation, and presentation of products and their related information. Besides its immediate search benefits, such annotations can also allow intelligent agents to perform their tasks better, as for example supporting agents aiming at comparing products or performing automatic trading.

More information
SWEPS website

Economic Sentiment Mining Based on Argumentation Structures

This project aims to investigate how information contained in argumentation structures can be harvested in order to refine and advance economic sentiment mining. Existing sentiment mining approaches enable analysis of a large number of text documents, mainly based on their statistical properties and possibly combined with numeric data. Most approaches are limited to simple word counts and largely ignore semantic and structural aspects of content. Yet, argumentation plays an important role in expressing and promoting an opinion. Therefore, in this project, models, methods, and algorithms that utilize argumentation structures for the automated discovery of sentiment in economic discourse are developed and applied.

FERNAT (Financial Events Recognition in News for Algorithmic Trading)

The FERNAT project focuses on the recognition of financial events in news messages, and applying these events in algorithmic trading. For this, a multi-disciplinary framework that uses techniques from finance, text mining, artificial intelligence, and the Semantic Web is developed. The framework consists of a news processing pipeline in which the identification of financial events in emerging news is automated while minimizing human intervention, and an algorithmic trading component where identified events are applied to trading algorithms.

More information
FERNAT website

TOWL (Time-Determined Ontology Web Language)

The TOWL project focuses on extending the OWL language with a temporal dimension in order to enable real-time context aware information analysis. Faster automatic analyses of corporate, market, and global news will become crucial in order to deliver equity trading performance. The usefulness of the TOWL language is demonstrated by implementing an application, aiming to support the daily decisions of stock brokers with a dynamic content environment.

More information
TOWL website

ABSTRACTE (Agent-Based Simulation of Trading Roles in Asynchronous Continuous Trading Environment)

The ABSTRACTE project focuses on modelling trading environments that are agent-based and experience a continuous asynchronous behaviour of participants. The proposed ABSTRACTE environment can be used to model several types of markets and an arbitrary number of trading strategies. The tool is designed with the aim to improve the study and understanding of market dynamics.

More information
PhD thesis by Katalin Boer-Sorbán, Agent-Based Simulation of Financial Markets: A Modular, Continuous-time Approach

Hera (Model-Driven Design of Web Information Systems)

The Hera project proposes model-driven design methodologies for developing Web Information Systems. Due to the diversity of Web users and their browsing platforms Hera aims at building personalized systems that provide the users with the desired information on their preferred browsing platform. Such systems are built in order to satisfy the information need of the users, as for example managers, anytime and anywhere.

More information
The Hera website

ViCore (Visualization of Concept Relations)

The ViCore project aims is to provide a comprehensive overview of a large amount of information by using data visualization techniques. By visualizing the concepts and their relationship one can, for example, understand the structure of an emerging research field or monitor the evolution of an established scientific domain. The ViCore project has resulted in a tool that allows the visualization of the scientific domains as, for example, the economic domain for the purpose of understanding the recent developments in this area.

More information
ViCore press release