Integration of Bagging and greedy forward selection on image pap smear classification using Naïve Bayes

摘要: Herlev dataset consists of 7 cervical cell classes, i.e. superficial squamous, intermediate columnar, mild dysplasia, moderate severe and carcinoma in situ is considered. The will be tested to classify two consisting normal abnormal cells. Seven different types classified separate the cells into classes which are 3 4 classes. There still some difficulties seven This Pap smear image has a class with number unbalanced Another condition that data features suspected irrelevant, so it difficult especially To handle imbalance, this study used ensemble method (Bagging). For handling had no contribution, we made feature selection Greedy Forward Selection. Furthermore, Naive Bayes was as learning algorithms. results obtained highest accuracy value for classification using model Selection 92.15%. As good enough Bagging 63.25% although needs improve.