IMPROVE FILTER METHOD FOR FEATURE SELECTION TECHNIQUE IN TEXT MINING
Keywords:Text mining, Feature selection, Filter method, wrapper method, embedded method, MFD algorithm
Huge amount data on the internet are in unstructured texts can’t simply be used for further
processing by computer, therefore specific processing method and algorithm is required to extract useful pattern.
Text mining is process to extract information from the unstructured data. A major difficulty of text classification
is high dimensionality of feature space. Feature selection method used for dimension reduction. We use filter
method for feature selection that use MFD algorithm but main drawback it is it necessary to assign value of F so
it’s work for only given features F we overcome this by proposed method in which F value select automatically by
pre analyses data set extract unique feature and create feature set F so its work for all unique features and also
reduce maximum redundancy by imposing threshold value.