A Bayesian Approach to the Data Description Problem

TitleA Bayesian Approach to the Data Description Problem
Publication TypeConference Paper
Year of Publication2012
AuthorsGhasemi, A., H. R. Rabiee, M. T. Manzuri, and M. H. Rohban
Conference NameAAAI
Date Published07/2012
Conference LocationToronto, Ontario, Canada
AbstractIn this paper, we address the problem of data description using a Bayesian framework. The goal of data descrip- tion is to draw a boundary around objects of a certain class of interest to discriminate that class from the rest of the feature space. Data description is also known as one-class learning and has a wide range of applications. The proposed approach uses a Bayesian framework to precisely compute the class boundary and therefore can utilize domain information in form of prior knowl- edge in the framework. It can also operate in the ker- nel space and therefore recognize arbitrary boundary shapes. Moreover, the proposed method can utilize un- labeled data in order to improve accuracy of discrimi- nation. We evaluate our method using various real-world datasets and compare it with other state of the art ap- proaches of data description. Experiments show promis- ing results and improved performance over other data description and one-class learning algorithms.