We utilize data mining and medical informatics techniques to enhance our understanding of risk factors for onset of disease or to enable the identification of novel predictors of key clinical outcomes.
We develop prediction models employing data mining technologies based on advanced classification algorithms (supervised learning), and apply methods such as decision trees and neural networks.
The development of these models takes into account both their quality and generalizability to determine the superior predictive model.
Our research in this field is focused on different steps of this process, such as:
- Panel creation and validation using feature selection and dimensionality reduction
- Data mining algorithms (supervised and unsupervised), such as decision trees, neural networks and cluster analysis
- Visualization and pattern evaluation