My research explores the interpretable application of statistics in the field of radiomics, a rapidly growing area of medical research, specifically in that of cancer. Radiomics is the process of mining large amounts of data from images and using the collected information to help make clinical decisions or to classify lesions, such as discriminating between malignant tumors and benign lesions. There are many different parts to this field, ranging from collecting the information to using it in statistical models. The feature sets collected are typically massive in size and scope, and my research both addresses the creation of new features to add to the radiomic feature space and building interpretable models for classification that take the reliability and robustness of the features in to account.
- Cremaschi, A., Argiento, R., Shoemaker, K., Peterson, C.B. and Vannucci M. (2019). Hierarchical Normalized Completely Random Measures for Robust Graphical Modeling. Bayesian Analysis, 14(4), 1271-1301.
- Shoemaker, K., Hobbs, B., Bharath, K., Ng, C., Baladandayuthapani, V. (2018). Tree-based Methods for Characterizing Tumor Density Heterogeneity. Pacific Symposium on Biocomputing. 23:216-227
2018 – “Hierarchical Normalized Completely Random Measures for Robust Graphical Modeling," Honorable Mention, ISBA Lindley Prize