@inproceedings{motta_hyperparameter_2019, abstract = {This paper describes a general pipeline for generating optimal vector representations of topological features of data for use with machine learning algorithms. This pipeline can be viewed as a costly black-box function defined over a complex configuration space, each point of which specifies both how features are generated and how predictive models are trained on those features. We propose using state-of-the-art Bayesian optimization algorithms to inform the choice of topological vectorization hyperparameters while simultaneously choosing learning model parameters. We demonstrate the need for and effectiveness of this pipeline using two difficult biological learning problems, and illustrate the nontrivial interactions between topological feature generation and learning model hyperparameters.}, author = {Motta, Francis and Tralie, Christopher and Bedini, Rossella and Bini, Fabiano and Bini, Gilberto and Eramian, Hamed and Gameiro, Marcio and Haase, Steve and Haddox, Hugh and Harer, John and Leiby, Nick and Marinozzi, Franco and Novotney, Scott and Rocklin, Gabe and Singer, Jed and Strickland, Devin and Vaughn, Matt}, booktitle = {2019 18th {IEEE} International Conference On Machine Learning And Applications ({ICMLA})}, date = {2019-12}, doi = {10.1109/ICMLA.2019.00185}, eventtitle = {2019 18th {IEEE} International Conference On Machine Learning And Applications ({ICMLA})}, keywords = {1 - Biology, 1 - Protein folding, 2 - Configuration space, 2 - Hyperparameter optimization, 2 - Machine learning, 2 - Persistence diagrams, 3 - Protein stability, 3 - Rocklin designs}, note = {{ISSN}: null}, pages = {1107--1114}, title = {Hyperparameter Optimization of Topological Features for Machine Learning Applications} }