Lead: Dr. Trevor Hastie
Statistical learning approaches allow us to extract insight from complex time-series data that varies widely in fidelity, sparsity, noisiness, and type.
Our team has demonstrated success applying statistical learning approaches to solve biomedical big data problems, but we’ve only scratched the surface. We will develop and validate statistical models for making predictions and identifying trends, correlations, and clusters in large-scale, sparse, and irregular time-varying measurements. We will analyze information from wearable sensors, smartphones, and clinical databases, in combination with the more traditional high-fidelity data collected in research labs. We will also develop automatic tools to generate insights from biomedical big data, streamlining the currently painstaking process of extracting surprising and plausible findings from inconsistent data in the literature and clinical or research databases.