Enhanced Artificial Bee Colony Algorithm for Liver Cancer Analysis
View Source
Lokanayaki K is an Associate Professor at School of Computer Science and Engineering, RV University. She holds Doctoral Degree (Ph.D.) in Computer Science from Bharathiar University, with research focusing on developing advanced methods for liver cancer analysis using ensemble sliding and swarm-based approaches. She has 18 years of teaching experience in various educational institutions. She has been involved in several research projects and has published extensively in international journals and conferences.Additionally, she has obtained patents in fields such as industrial cyber-physical systems and cognitive internet of things. She is also a member of multiple professional bodies and has attended numerous workshops and faculty development programs to enhance her teaching and research skill
Success is not final, failure is not fatal: It is the courage to continue that counts.
ABSTRACT During the development of computer technology, computer-aided diagnosis (CAD) technology, used in quantitative analysis of medical imaging, arose at a historic moment and became a research hotspot in medical imaging. Discrimination of hepatocellular carcinoma (HCC) in the liver is a challenge in the histopathologic diagnostics. For this reason, there is an urgent need for new detection methods to improve the accuracy of the diagnosis of liver cancer. Traditional machine-learning approaches are neural network (NN)-based. Cost-sensitive learning and a support vector machine (SVM) is observed to provide a good result in the case of balanced data sets; however, it is not capable of dealing with the classification of imbalanced data sets. These machine-learning approaches may be biased toward the majority class, thus producing a poor predictive accuracy over the minority class. In this paper, a novel technique for the purpose of liver cancer cell classification and root liver cancer cell recognition is proposed. The objective is to automatically categorize several classes of liver cancer cells and to discover the root cancer cell. To solve this problem, initially, preprocessing on noisy imbalanced data sets is carried out by means of improved weighted synthetic minority oversampling technique (IWSMOTE)-based oversampling and evolutionary undersampling. An ensemble-based learning algorithm (DataBoost.IM) with SVM is employed for final classification to classify the cancer cells and non- cancer cells. Finally, the enhanced artificial bee colony (EABC) clustering is applied to discover the root cancer cell. The proposed EABC clustering approach is tested using the liver cancer cell data set, providing an accuracy level of 96.15 %, which is 95.61 % and 92.80 % higher than the ant colony optimization (ACO) and artificial bee colony (ABC) algorithms, respectively.
