Classification with incomplete data and ensemble learners for the prediction of cervical cancer risk

Abstract

Incomplete data is an important problem in analyzing medical data sets. In this study, a comparative analysis of ensemble learning algorithms was carried out for the prediction of cervical cancer risk with incomplete data. Cervical cancer is one of the most common cancers for women world-wide, and many researchers focused on this disease. The dataset was collected from UCI Machine Learning Repository. Mean imputation was used to deal with missing values and some ensemble and standalone classifiers were used to analyze the dataset for the evaluation of the performance. This study supported that imputation approaches and ensemble learning can improve the performance of learning algorithms. © 2018 Association for Computing Machinery.