An Optimized RBF-Neural Network for Breast Cancer Classification

Abstract

This paper introduces an optimized RBF-Neural Network for breast cancer classification. The study is based on the optimization of the network through three learning phases. In the first phase, K-means clustering method is used to define RBFs centers. In the second phase, Particle Swarm Optimization is used to optimize RBFs widths. In this phase, a pseudo inverse solution is used to calculate the output weights. Finally, in the third phase, the back-propagation algorithm is used for fine-tuning the obtained parameters, namely centers, widths and output weights. The back-propagation is then initialized with the obtained parameters instead of a random initialization. To evaluate the performance of the proposed method, tests were performed using the Wisconsin Diagnostic Breast Cancer database. The proposed system was compared with a network trained only with BP and a network trained with K-means + PSO. The results obtained are promising compared to other advanced methods and the proposed learning method gives better results by combining these three methods.

Keywords

• K-means,Radial Basis Function Networks,Classification,Neural Networks,Particle Swarm Optimization

References

1. Reference1 Eberhart, R., Kennedy, J.: A new optimizer using particle swarm theory. In: Micro Machine and Human Science, 1995. MHS’95., Proceedings of the Sixth International Symposium on. pp. 39–43. IEEE (1995)
2. Reference2 Kennedy, J., Eberhart, R.: Particle swarm optimization, proceedings of ieee international conference on neural networks (icnn’95) in (1995)
3. Reference3 Lichman, M.: Uci machine learning repository. irvine, ca: University of california, school of information and computer science. URL http://archive. ics. uci. edu/ml (2013)
4. Reference4 Mert, A., Kılıc¸, N., Bilgili, E., Akan, A.: Breast cancer detection with reduced feature set. Computational and mathematical methods in medicine 2015 (2015)
5. Reference5 Qasem, S., Shamsuddin, S.: Generalization improvement of radial basis function network based on multi-objective particle swarm optimization. J. Artif. Intell 3(1), 1–16 (2010)
6. Reference6 Rumelhart, D., Hinton, G., Williams, R.: Learning internal representation by error propagation, parallel distributed processing: exploration in the microstructure of cognition. MIT press Cambridge pp. 318–362 (1986)
7. Reference7 Venkatesan, A.S., Parthiban, L.: A novel optimized adaptive learning approach of rbf on biomedical data sets. Research Journal of Applied Sciences, Engineering and Technology 8(4), 541–547 (2014)
8. Reference8 Zarbakhsh, P., Addeh, A., et al.: Breast cancer tumor type recognition using graph feature selection technique and radial basis function neural network with optimal structure. Journal of cancer research and therapeutics 14(3), 625 (2018)