Breast cancer detection using neural network LVQ on breast cancer database

Introduction & Aim: One of the main problems and unresolved in the successful treatment of cancer, early diagnosis and lack of proper method in it. Hence the use of data mining models can lead to increased diagnostic accuracy of early diagnosis of cancer. Methods: The study is fundamental analysis. The database used in this study. WDBC was used in UCI database, the database contains 569 records of FNA sample with 31 features. Based on the features included in this data set, LVQ neural network diagnostic model were created. Models created with cross-validation method were evaluated in the software Matlab 2013a. And the 10 model, a model that had the highest efficiency was selected as the final model. Results: The model parameters based on the highest performance standards in the harvest were as follows: sensitivity: 98.1%, specificity: 96.2%, accuracy: 94.3 Performance: 0.98.8. Conclusion: The results of this study showed that using an artificial intelligence algorithm and its application in modeling can be very sensitive to the method of medical screening, breast cancer diagnosed. In fact, using these methods to help doctors design new systems that facilitate the diagnostic processes and treatment