Author : Ajay Singh Mavai and Sadhna K. Mishra

Social media is very useful medium for data collection and predicting status and situation of a user. Many real world applications and data sets often contain many missing elements which may present a big obstacle for many learning algorithms, which usually require a complete data set to build the model. Most algorithms that automatically develop a rule based model are not well suited to deal with incomplete data, till now many missing value estimator system implemented using many techniques such as Bootstrap, K- nearest neighbor, Support Vector Machine, Bayesian Classification, Rough set approach and by using combination of different classifiers, hybrid approach as well as layered approach. In this paper we describe the comparative study of different Data Mining techniques to implement a Missing Value Estimator System (MVES) and also describe the techniques and methods for data selection, finding missing values and modeling to sort the useful information which could be used in other application in different ways.

Keywords: Missing Value Estimator System (MVES), Social Media(SM), Data Mining, KNN, Bootstrap, SVM.

Article published in International Journal of Current  Engineering  and Technology, Vol.4,No.4 (Aug- 2014)