In predictive modeling, the typical paradigm consists of learning a single model from a database of patient cases, which is then applied to predict outcomes for any future patient. Such a model is called a population-wide model because it is intended to be applied to an entire population of future cases. In contrast, patent-specific modeling focuses on learning models consists of learning models that are personalized to the characteristics of the patient at hand. Our work focusses on patient-specific models that are optimized to perform well on a specific patient are likely to be more precise than the typical population-wide models that are optimized to have good predictive performance on average on all future patients.
Research Projects and Collaborations
Automated visual analytics combines visual analytics with automated analysis for discovery of patterns in data. Visual analytics is interactive analysis facilitated by interactive visual interfaces where the domain expert interacts with the data visually to identify interesting patterns. However, it is effort-intensive and is readily applicable only to datasets with low dimensionality and small sample sizes. Our work focusses on automated visual analytics that combines visualization and automated analysis methods to take advantage of the rapid search that automated methods provide with the ability to identify novel and rich patterns that visual analytics provides.
Personalized medicine calls for the use of clinical, genomic and environmental data to more precisely evaluate risk, diagnose, assess prognosis, and tailor therapies to the individual. Genomic medicine is driving personalized medicine and focuses on the use of information obtained from sequences such as whole exomes and whole genomes. Genomic information, in combination with other clinical data, will lead to increased understanding of the biology of human health and disease, improved prediction of disease and effect of therapy, and ultimately the realization of precision medicine. Our work focuses on single nucleotide variants (SNVs) data obtained from genome-wide studies (GAWSs), and more recently whole exomes.
Electronic medical records (EMRs) are capturing increasing amounts of patient data that can be leveraged by machine-learning methods for computerized decision support. Our work focusses on developing and implementing machine-learning methods for detecting adverse drug events and for identifying anomalies or deviations in therapy and clinical management of patients.
Increasing amounts of data in EMRs pose challenges to aggregate, synthesize, and identify patterns for clinical care. Our work focusses on the development of adaptive and learning components in EMRs that will provide decision support using the right data, at the right time.