An ensemble approach to improved prediction from multitype data - Statistics > ApplicationsReportar como inadecuado




An ensemble approach to improved prediction from multitype data - Statistics > Applications - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Abstract: We have developed a strategy for the analysis of newly available binary datato improve outcome predictions based on existing data binary or non-binary.Our strategy involves two modeling approaches for the newly available data, onecombining binary covariate selection via LASSO with logistic regression and onebased on logic trees. The results of these models are then compared to theresults of a model based on existing data with the objective of combining modelresults to achieve the most accurate predictions. The combination of modelpredictions is aided by the use of support vector machines to identifysubspaces of the covariate space in which specific models lead to successfulpredictions. We demonstrate our approach in the analysis of single nucleotidepolymorphism SNP data and traditional clinical risk factors for theprediction of coronary heart disease.



Autor: Jennifer Clarke, David Seo

Fuente: https://arxiv.org/







Documentos relacionados