A model for predicting outfit sales: Using data mining methods

Abstract

The objective of this study is the development of a model that will use data mining techniques to predict the sum of outfit sales. Data mining is the method of extracting appropriate information from the dataset and converts them into a valuable one. In this research, I have adopted the data mining procedures such as preprocessing (e.g., Data cleaning, missing values replacement and reduction of Data, etc.) on the data of outfit sales, which enhance the performance up to 20%. I have then train the different classifiers, for example, Naive Bayes, Multilayer Perceptron, and K-star. Finally, assessed the performance of every classifier via tenfold cross validation and compared the outcomes, where, Multilayer perception classifier showed the highest performance with 83.8% accuracy and Nave Bayes showed the lowest performance with 71.6% accuracy. This research also showed the standard level of improvement in the performance (up to 9%), as compared to the other predicted models of outfit sales. The total research was implemented in a Data mining software tool, WEKA.