Analysis of Unsupervised Dicretization Methods Impact on C4.5 Classification
In this paper five unsupervised methods are considered to study the effect on C4.5 classification. For studying the impact of dicretization process twenty three data sets are downloaded from the UCI Machine learning repositories and Keel repositories. R language is used to preprocess and the weka tool is used to apply dicretization on datasets. The Equal Width and Equal Frequency procedures are considered for five methods and C4.5 dicretization. The average efficiency is consider as performance measure for comparative study .The five dicretization methods are implemented in R Studio and the weka tool is used for executing C4.5 classification with all default parameters. The experimental results are validated using statistical results. The results are presented in tables as well as graphs and conclusions are drawn.