Call For Papers
Contact Us

  Diagnosis of Breast Cancer by Combining the Techniques of Data Mining and Artificial Immune System  
  Authors : Esmat Banihashem; Touraj Banirostam
  Cite as:


Breast cancer is known as the most common cancer among women so that in 2012, 29% of cases were diagnosed among women have been infected with breast cancer. Early diagnosis of breast cancer (max. 5 years after the first cell division of the cancer) the patient's chance of survival increases from 56% to over 86% .Therefore existence of a precise and reliable system for timely diagnosis of benign and malignant breast tumors is very important. A lot of details about cancer characteristics make diagnosis difficult for doctors, Therefore, data analysis methodologies will be a useful assistant for doctors to diagnose cancer. Currently, using FNA as a method for tumor mass sampling and testing on that type of tumor (benign and malignant) is indicated. By performing data mining algorithms on the data obtained from the sampling, a higher accuracy can be detected. The data set used in this study was extracted from the data set in the machine learning tank of university of California known as UCI. In this thesis we want to benign or malignant breast cancer detection used from new method iLA-VQIS (combination of two competitive algorithm : LVQ and evolutionary immune system algorithm) to improve detection. In fact, the training of neural network weights is done using an artificial immunological clonal algorithm. Inside this function, instead of gradients based on the neural network, evolutionary optimization method will be used and usually the function arguments will not be changed. Point of subscription between artificial immune system evolutionary optimization algorithms with the neural networks topic, after designing the network structure is the learning process that ends with an optimization problem and finally the results are evaluated with three criteria for precision, accuracy and recall. Simulation operations performed in MATLAB and results of the proposed LA-VQIS algorithm has been compared with basic algorithms such as Kohenen LVQ algorithm, combined decision tree algorithm with genetic algorithm, K_SWM algorithm, SWM and MSAIS algorithm under the same conditions and based on the correctness and accuracy of the diagnosis.


Published In : IJCSN Journal Volume 6, Issue 5

Date of Publication : October2017

Pages : 539-546

Figures :10

Tables : 06


Esmat Banihashem : Collegian of Master of Science in Artificial Intelligence, Electronic Unit of Azad University Tehran, Ir Iran.

Touraj Banirostam : Department of Computer Engineering, Islamic Azad University, Central Tehran Branch, Tehran, Iran.


breast cancer, data mining, LVQ neural network, artificial immune system

In the current research, we focus on the special application of synthetic immune algorithm. In fact, we investigated the data mining process through the synthetic immune function. The main part of this research is investigating the neural network education by means of the syntactic immune algorithm. In other words, by the supervising the syntactic immune program in the Train neural network, the educated data transforms to the antibody environment and the remained data are transformed to the antigen environment. Finally, three criteria would be evaluated as the following: accuracy, supervise and precision. Totally, the results of evaluated simulation A-VQIS with 97.8% accuracy illustrate the effect of LA-VQIS method on the data in order to diagnosis the breast cancer. Thus, we implemented the knowledge of a doctor as the smart system.


[1] A. Abbas and A. Lichtman, Basic Immunology, Updated Ed. 2006-2007, W.B. Saunders Company, 2007. [2] P. Reichardt and M. Gunzer, "The Biophysics of T Lymphocyte Activation In Vitro and In Vivo," in Cell Communication in Nervous and Immune System. vol. 43: Springer Berlin / Heidelberg, 2006, pp. 199-218. [3] F. M. Burnett, "The Clonal Selection Theory of Acquired Immunity", Cambridge University Press, (1959). [4] Kuo.LungWua, Miin.ShenYangb," Alternative learning vector quantization" , Pattern Recognition 351 – 362,2006. [5] E. de Bodt, M. Cottrell, P. Letremy, M. Verleysen, "On the use of selforganizing maps to accelerate vector quantization", Neurocomputing 56 (2004) 187–203. [6] K.L. Wu, M.S. Yang, "A fuzzy-soft learning vector quantization",Neurocomputing 55 (2003) 681–697. [7] Bichen Zheng, Sang Won Yoon, Sarah S. Lam, “Breast cancer diagnosis based on feature extraction using a hybrid of K-means and support vector machine algorithms”, Expert Systems with Applications 41 (2014) 1476–1482. [8] Lingxi Peng, Wenbin Chen, Wubai Zhou, Fufang Li, Jin Yang, Jiandong Zhang, An immune-inspired semisupervised algorithm for breast cancer diagnosis, Computer Methods and Programs in Biomedicine (2016), http://dx.doi.org/doi: 10.1016/j.cmpb.2016.07.020. [9] Zheng, B., S.W. Yoon, and S.S. Lam, Breast cancer diagnosis based on feature extraction using a hybrid of K-means and support vector machine algorithms. Expert Systems with Applications, 2014. 41(4): p. 1476-1482. [10] Asgari. Somayye , Study of Data Ming in Diagnosis of Breast Cancer , 3RD Electronic Conference of New Articles on Technology Science , 2015.