Our Methods

 

The MezeyLab develops computational analysis methods from the fields of Computational Statistics and Machine Learning - fields that often describe closely related approaches in different languages. Our focus is on understanding why particular methods work well for analyzing a given big biological data type for a specific application, whether the objective is to return actionable insights or to produce high value predictions. Our overall development process starts by understanding what outcomes could be achieved in theory and then careful empirical assessment of what can be achieved in practice, where a common outcome is we often find widely applied methods do not work as well advertised for a given application. Our work developing computational methodology for analyzing big biological data types produces output in a number of areas, including theory, extensions of existing methods, development of new methods, derivation of algorithms and development of frameworks for implementation. Please the see our publications and the following pages for representative examples of our computational methodology work including: