“I seek to discover and model information in texts using text mining and AI methods.”

I am assistant professor in Computer Science at Utrecht University and member of the Data Intensive Systems group. My principal expertise lies in the area of natural language processing and in particular in extracting and modelling information in written texts.

Research interests:  

• Text mining & information retrieval research for information management & decision support
• Interpretability of text mining methods
• Text information processing applications (business, industry, health & digital humanities)
• Knowledge representation (ontology development, lexical knowledge resources, metadata)
• Data integration, linked data & semantic web technologies

I find human language fascinating in all its forms, both as a communication medium in a cultural and societal context, and as a complex information and knowledge representation system. My main research focus lies in the analysis of text for information acquisition and management, considering information retrieval, keyword and domain-specific terminology extraction, document classification, topic modelling, semantic entity classification, relation extraction, distributional semantics and large language models.

My research focuses on real-world, raw operational data to derive insights and support strategic decisions in a variety of domains, such as business, social sciences, digital history and healthcare. Real-world data is very diverse, consisting of large collections of numeric data and measurements, domain-specific structured data in schemas and business rules, and large amounts of unstructured textual data. In real-life language processing applications, important resources such as annotated training data are often not available and we often deal with language varieties that present particular challenges (OCR errors, multi-linguality, availability, personal data protection), such as historical or technical/healthcare language varieties.

At early stages of my research I worked on rule-based and knowledge-based systems for the processing of texts. Afterwards, I have been experimenting with unsupervised and/or data-driven methods while also considering issues related to data curation and normalisation, (meta-)data enrichment, linking and integration, and text data modelling in knowledge resources.

My latest research focuses on text mining electronic health records for information discovery, classification and prognosis.