Data Science is a relatively new interdisciplinary field, taking concepts from statistics and machine learning to produce predictive models. However, Systems Science concepts (such as feature-feature interactions and dynamics) have been largely underutilized in Data Science. In this talk, I'd like to start a discussion of specific ways that Systems Science can inform Data Science. I will start with examples of network analysis in my research that have led to better predictive models, and end with a discussion about the interpretability of black box predictors such as neural networks. I believe that Systems Science approaches can enhance Data Science by providing a deeper understanding of interactions between features and interpretability.
I am an Assistant Professor in the Division of Bioinformatics and Computational Biology in the Department of Medical Informatics and Clinical Epidemiology at OHSU (BCB/DMICE). My research focus is on the Systems Biology of Complex Diseases, especially within cancer. I use integrative modeling approaches (such as network analysis and graphical models) across OMICs types to achieve this. I am also an active participant in the Portland Data Science community, especially the R programming community. More information at http://laderast.github.io/
Big data, System theory, System analysis, Neural networks, Machine learning
Computer Sciences | Data Science | Medicine and Health Sciences
Laderas, Ted, "How are Data Science and Systems Science Connected?" (2018). Systems Science Friday Noon Seminar Series. 81.