NOVEL MULTIPLE IMPUTATION COMPARISON WITH SIMPLE LINEAR CLASSIFIER, SUPPORT VECTOR MACHINE AND NAÏVE BAYES CLASSIFIER

##plugins.themes.bootstrap3.article.main##

Nithya Rani A Antony Selvdoss Davamani

Abstract

Incomplete data is a common obstacle to the analysis of data in a variety of fields, ranging from clinical trials to social sciences. Missing values can occur for several different reasons including failure to answer a survey question, dropout, planned missing values, intermittent missed measurements, latent variables, and  quipment malfunction. Multiple imputations is one method for handling incomplete data that accounts for the variability of the incomplete data. This procedure does so by filling in plausible values several times to create several complete data sets and then appropriately combining complete data estimates using specific combining rules. We introduced the methodology of multiple imputations in multiple stages and the associated comparison of Simple Linear Classifier, Support Vector Machine and Naïve Bayes Classifiers needed for implementation. We demonstrated via simulations that we have an efficient estimator under the assumption.

##plugins.themes.bootstrap3.article.details##

Section
Articles