Multimodal data mining involves analyzing more than one form of data to discover hidden patterns. It can support advances in a variety of fields, from psychology to education to health care and more. As Canada Research Chair in Multimodal Data Mining, Dr. Belkacem Chikhaoui is working on multimodal and heterogeneous big data fusion, analysis and mining to build a new mathematical framework and a solid foundation for applications with potentially significant social impacts.
He and his research team are mining multimodal heterogeneous big data and discovering causal relationships and hidden patterns for predictive analytics. They are also developing new algorithms by combining manifold learning models with causality theory and opportunities offered by new big data processing solutions. They are then applying these algorithms to solve real-world problems—such as anticipating violent behaviour, predicting student drop-out rates, and detecting urinary incontinence in residents of long-term-care facilities.