文章基本信息

标题：Active Clustering based Classification for Cost Effective Prediction in few Labeled Data Problem
本地全文：下载
作者：Gábor SZŰCS ; Zsuzsanna HENK
期刊名称：Economy Informatics
印刷版ISSN：1582-7941
出版年度：2015
卷号：15
期号：1
页码：5
出版社：INFOREC Association
摘要：In many data mining problems related to business it is hard to obtain labeled instances. Whenthe labeled data set is not large enough the classifiers often perform poor results. Neverthelesssemi-supervised learning algorithms, e.g. clustering based classification can learn fromboth labeled and unlabeled instances. We have planned and implemented a semi-supervisedlearning technique by combining the clustering based classification system with active learning.Our active clustering based classification method first clusters both the labeled and unlabeleddata with the guidance of labeled instances, then queries the label of the most informativeinstances in an active learning cycle and after that classifies the data set. At costbenefit analysis comparing the results of our system with the supervised learning and clusteringbased classification it can be concluded that our solution saves the largest cost.
关键词：Active Learning; Data Mining; Semi-Supervised Learning; Clustering Based;Classification; Cost Benefit Analysis