Feature Selection, Ranking of Each Feature and Classification for the Diagnosis of Community Acquired Legionella Pneumonia

摘要: Diagnosis of community acquired legionella pneumonia (CALP) is currently performed by means laboratory techniques which may delay diagnosis several hours. To determine whether ANN can categorize CALP and non-legionella community-acquired (NLCAP) be standard for use clinicians, we prospectively studied 203 patients with (CAP) diagnosed tests. Twenty one clinical analytical variables were recorded to train a neural net two classes (LCAP or NLCAP class). In this paper deal the problem diagnosis, feature selection, ranking features as function their classification importance, design classifier criteria maximizing ROC (Receiving operating characteristics) area, gives good trade-off between true positives false negatives. order guarantee validity statistics; train-validation-test databases rotated jackknife technique, multistarting procedure was done in make system insensitive local maxima.