Title :
Bootstrapping nonparametric feature selection algorithms for mining small data sets
Abstract :
This paper presents feature selection algorithms based on nonparametric feature ranking indices, and demonstrates for small data sets that by bootstrapping feature ranking indices one uniformly (over various data sets and different ranking indices) improves the performance of correct detection of true features (i.e., probability of the top ranking features matching the true ones)
Keywords :
computer bootstrapping; data mining; learning (artificial intelligence); neural nets; probability; bootstrapping; data mining; feature ranking indices; neural nets; nonparametric feature selection; probability; small data sets; Computer vision; Data analysis; Data mining; Distribution functions; Input variables; Life estimation; Neural networks; Scattering; Testing; Training data;
Conference_Titel :
Neural Networks, 1999. IJCNN '99. International Joint Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-5529-6
DOI :
10.1109/IJCNN.1999.833470
