Improving The Accuracy Of Neural Network Technique with Genetic Algorithm for Cervical Cancer Prediction

• Published in: 2018 6th International Conference on Cyber and IT Service Management (CITSM) pp. 1 - 7
• Main Authors: Brawijaya, Herlambang, Widodo, Slamet, Samudi
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2018
• Subjects: Biological neural networks; Cervical cancer; cervical cancer prediction; genetic algorithm; Genetic algorithms; neural network; Sociology; Statistics; Training
• DOI: 10.1109/CITSM.2018.8674298

New cases of cervical cancer in Indonesia more than 13,700 and nearly 7500 deaths caused by this cancer in 2008. Cervical cancer ranked third disease experienced by women around the world and ranked fourth cause of cancer deaths. In 2008, there were 529,800 women diagnosed with cervical cancer and 275,100 of them died from the disease. In this research, we tested the model using neural network with neural network based on genetic algorithm to predict cervical cancer. Several experiments were conducted to obtain optimal architecture and produce more accurate prediction accuracy. Experimental results with various combinations of research parameters showed that experiments using neural network obtained the best accuracy value is 94.51% with AUC value 0.961. While experiments using genetic algorithm based neural network obtained 96.26% accuracy value with AUC value 0.968. The results show that the model formed by neural network based on genetic algorithm produces better accuracy compared to neural network without genetic algorithm. Based on the tests that have been done can be concluded that the genetic algorithm can improve neural network performance in predicting data of cervical cancer.