Abstract: Classification is undoubtedly gaining major importance in the fields of machine learning, pattern recognition genetic engineering and bio medical sciences where it can be used for automated decision making. Mostly these areas contain datasets having large number of dimensions which require some preprocessing. Thus dimension reduction is a preprocessing step carried out prior to classification so that the classifiers can be designed in an easy to compute way. However while doing so, it must also retain the accuracy and must not lead to loss of information. Thus effective classification must be carried out by employing proper dimension reduction techniques. The paper discusses in brief about the dimension reduction techniques. It also describes the system developed for dimension reduction and use of the tool WEKA for dimension reduction and preprocessing. Finally a comparative study of the results obtained by the system and WEKA is done.