Learning the Attribute Selection Measures for Decision Tree

Publisher:
SPIE
Publication Type:
Conference Proceeding
Citation:
Fifth International Conference on Machine Vision (ICMV 2012): Algorithms, Pattern Recognition, and Basic Technologies, 2013, 8784
Issue Date:
2013-03-13
Full metadata record
Files in This Item:
Filename Description Size
87842S.pdfPublished version264.58 kB
Adobe PDF
Decision tree has most widely used for classification. However the main influence of decision tree classification performance is attribute selection problem. The paper considers a number of different attribute selection measures and experimentally examines their behavior in classification. The results show that the choice of measure doesn’t affect the classification accuracy, but the size of the tree is influenced significantly. The main effect of the new attribute selection measures which base on normal gain and distance is that they generate smaller trees than traditional attribute selection measures.
Please use this identifier to cite or link to this item: