An Extensive Benchmark Experimental Evaluation of Methods for Multi Label Learning In R

Abstract

A smart product is one that is able to immingle with masses. Sensible merchandise does not seem to be solely easy merchandise, however, with a touch of cleverness supplemental to permit the user some flexibility operative. Smart product adapts to the context of the user and does not force the user to adapt to that. Sensible merchandise have a group of properties that creates them distinctive area unit self informative, self organizing, extensible, self property, device capabilities, practicality, integrity, user services, property. The client’s ranking or priority whereas shopping for varied sensible merchandise area unit dynamical day by day as a result of advancements in technology and customer principally target the advanced options of the sensible merchandise they are shopping for. This paper principally shows, however, affectively sensible merchandise area unit utilized by the shoppers and area unit hierarchic based mostly upon their performance by exploitation R language and WEKA. By using R we can have a deep analysis over the various smart products and the user can be able to know the most widely purchased smart products according to their ranking. We can have deep analysis of smart products using data mining classification and prediction techniques such as J48, Random Forest machine learning algorithms in WEKA and R Language. R allows wide number of smart products data and analyzes with in limited resources. The WEKA and R language is opted to see the classification and prediction performances. Four measures (sensitivity, specificity, accuracy, F–measure) of performance here considered are based on confusion matrix/Error Matrix of R and WEKA, table of counts revealing the performance of each algorithm’s confusion regarding the true classifications and predictions. The observation of all the four performance measures used to analyze the smart products use.