Inferring probability of guessing from item response data using bayes' theorem

Abstract

Outlier detection is a primary step in many data mining applications. Outlier means a marked response data correctly by guessing in item response data. Guessing an answer or a judgment about something without being sure of all the facts act as a noise in data mining. It is important to clean noise data for producing good results in data mining. In order to clean noise data, it is needed to detect correct answers marked by guessing among item response data. In this paper, we present a Bayesian approach to infer a probability of guessing for items. © Springer-Verlag Berlin Heidelberg 2014.