Title: Ensemble feature selection and deep learning ensemble classifier for cervical cancer diagnosis

Authors: Anjali Kuruvilla; B. Jayanthi

Addresses: School of Computer Studies, Rathnavel Subramaniam College of Arts and Science, Coimbatore, Tamil Nadu, India ' School of Computer Studies, Rathnavel Subramaniam College of Arts and Science, Coimbatore, Tamil Nadu, India

Abstract: Cervical cancer has become one of the foremost causes of cancer mortality in women. Developing a new approach requires improving system performance precision. Cervical cancer has many risk factors. Cervical cancer test parameters must be considered when classifying patients based on results. Recently, cervical cancer prediction features have been assessed using several feature selection methodologies. Ensemble feature selection (EFS) outperforms individual techniques. This study classifies cervical cancer cells using a deep learning ensemble (DLE) classifier. EFS combining the results of EBFO, EEHO, and RFE yield better results than using a single FS approach. DLE classifier uses heterogeneous base learners (GAN, BGRU, and DWCNN) and a meta-learner to predict cervical cancer from risk variables. DLE classifier builds numerous basic classifiers on which a new predictor outperforms any component. Stacking trains in many models on one dataset. The model is generated by segmenting the training set again using K-fold cross-validation. The suggested system uses DLE classifier and synthetic minority oversampling technique (SMOTE). The data source has 32 features and four classes: Hinselmann, Schiller, cytology, and biopsy. Precision, recall/sensitivity, F-measure, specificity, and accuracy are calculated using a confusion matrix to determine the superiority of all classification algorithms like random forest (RF) and GAN.

Keywords: bidirectional gated recurrent unit; BGRU; cervical cancer; entropy butterfly optimisation algorithm; EBFO; ensemble feature selection; EFS; dynamic weight convolutional neural network; DWCNN; synthetic minority oversampling technique; SMOTE; deep learning ensemble; DLE.

DOI: 10.1504/IJBRA.2023.139123

International Journal of Bioinformatics Research and Applications, 2023 Vol.19 No.5/6, pp.430 - 461

Received: 19 Aug 2023
Accepted: 30 Oct 2023

Published online: 14 Jun 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article