Recommendation of outfit helps the people in taking the right decision while purchasing and also increases the sales. The analysis of the accuracy of the classified dataset using various data mining techniques and algorithms is the key concept of this paper. The accuracy when the algorithms are applied on the balanced dataset, imbalanced dataset, dataset with attribute reduction and without attribute reduction is compared. To perform the attribute reduction, we are using cfsSubsetEval, consistencySubsetEval and chisquaredAttributeEval. The algorithms that are used to classify the dataset are Random Forest, Naive Bayes, zeroR, Multilayer Perception, RBF Network and AdaboostM1. The main challenge is thatthe virtual dataset is imbalanced through which we got poor results with less accuracies. This dataset is balanced using SMOTE analysis to obtain higher accuracies and also attribute reduction is performed to compare the accuracies obtained. In comparison with the existing method, the maximum accuracy rate produced by the Poonkuzhali Sugumaran and Vinodh Kumar Sukumaran  was 98% using hybrid classifier ID3 and AdaBoost algorithms. In the proposed method, the dataset when balanced by SMOTE analysis and classified by Random Forest algorithm, it results in 99.86% of accuracy in recommending the outfit. © 2020, Research Trend. All rights reserved.