Abstract

Feature Selection is a process of selecting a subset of relevant features from a huge dataset that satisfy method dependent criteria and thus minimize the cardinality and ensure that the accuracy and precision is not affected ,hence approximating the original class distribution of data from a given set of selected features. Feature selection and feature extraction are the two problems that we face when we want to select the best and important attributes from a given dataset Feature selection is a step in data mining that is done prior to other steps and is found to be very useful and effective in removing unimportant attributes so that the storage efficiency and accuracy of the dataset can be increased. From a huge pool of data available we want to extract useful and relevant information. The problem is not the unavailability of data, it is the quality of data that we lack in. We have Rough Sets Theory which is very useful in extracting relevant attributes and help to increase the importance of the information system we have. Rough set theory works on the principle of classifying similar objects into classes with respect to some features and those features may collectively and shortly be termed as reducts.