Beta Version Available!
ECM: An Exposure Continuum Mapping framework for integrating spatially-correlated learning into study of environmental mixtures
The ‘exposure continuum map (ECM)’ framework seeks to enhance the study of complex chemical mixtures with by integrating intuitive clustering algorithms with novel spatially-correlated learning approaches to improve inferences. Conceptually, spatially-correlated learning improves statistical inferences by incorporating information from neighboring features in effort to improve estimation. In effect, this pooling of information may help compensate for limited sample size (e.g., rare exposures) and scenarios with high variation (e.g., outliers) often observed in environmental mixtures studies. Such strategies have a long and successful history outside of geography as a broad range of fields from text mining to cognitive neuroscience have realized the benefits of adopting such principles into the study of complex relationships