Information Processing

Development of algorithms to generate information and knowledge optimally

The research group Information Processing concerns itself with the automated extraction and processing of information from data of many different natures. To do this, techniques from the areas of machine learning, natural language processing, data mining and statistics are combined. The intent is to analyse known data as precise as possible to gather new insights. End users profit from accurate estimations and predictions for unknown data, but also from products that can learn from individual usage.

Research activities

Modelling, simulating and optimizing processes and organisations
Implementation of methods to compare plans with their implementation, improving the control of the process and supporting decision-making
Development and evaluation of algorithms for optimal information generation
Data collection, modelling and interpretation of social networks

Research topics

Machine Learning
Natural Language Processing
Algorithmic and combinational optimization methods
Social Network Analysis
Data Mining
Information Retrieval / Ontologies

Competencies

Optimization

Through Optimization the lifecycle of a building or the performance of a construction project among others are simulated in the computer and optimized. The process modelling enables the comparison between the plan and the actual realization, improves the process control and supports the decision making.

Social Network Analysis

The use of social networks on the internet is an everyday activity for many people. In the last years many applications appeared that enable easy publication and networking on the internet. Social network analysis helps to better understand complex relations between the actors. The insights resulting from the analysis can be used to consult organisations, improve development processes in organisations and make forecasts.

Selected Projects

Yooture

The School of Engineering FHNW has developed a sophisticated matching method for the innovative start-up Yooture. More...

Snappit

The goal is to implement a recommender system for the first reverse-auction platform of Switzerland. By learning from the behaviour data of all users, personalized recommendations of auctions will be provided to the user.

Kennwerte

Thanks to machine learning, Kennwerte.ch is able to estimate the price of real estates with very little input necessary. Export knowledge by architects and building estimators is combined with data analysis and statistics.

DrugSafety

The goal of the DrugSafety project is to achieve a partial automation of the reporting of drug-induced adverse reactions. Various information from medical reports is automatically extracted and processed.

Kontakt

Prof. Dr. André Csillaghy