Using Five Machine Learning for Breast Cancer Biopsy Predictions Based on Mammographic Diagnosis

Abstract

Breast cancer is one of the causes of female death in the world. Mammography  is  commonly  used for  distinguishing  malignant  tumors  from benign  ones. In  this research,  a mammographic  diagnostic  method  is  presented  for breast  cancer  biopsy  outcome  predictions  using  five machine learning which includes: Logistic Regression(LR), Linear Discriminant Analysis(LDA), Quadratic Discriminant Analysis(QDA), Random Forest(RF) and Support  Vector Machine(SVM)  classification.  The testing  results showed  that  SVM  learning  classification  performs better than other with accuracy  of  95.8%  in  diagnosing  both malignant and benign breast cancer,  a  sensitivity  of  98.4%  in  diagnosing    disease,  a specificity of 94.4%.

 Furthermore, an estimated area of the receiver operating characteristic  (ROC)  curve  analysis for Support vector machine (SVM) was  99.9%  for  breast  cancer outcome  predictions, outperformed  the  diagnostic  accuracies  of  Logistic Regression(LR), Linear Discriminant Analysis(LDA), Quadratic Discriminant Analysis(QDA), Random Forest(RF)    methods.  Therefore,  Support Vector Machine (SVM)  learning classification  with  mammography  can  provide  highly  accurate and consistent diagnoses in distinguishing malignant and benign cases for breast cancer predictions.

Keywords

• Logistic Regression,Linear Discriminant Analysis,Random Forest,Quantitative Discriminant Analysis,Support Vector Machine

References

1. Department of Health and Human Services Centers for Disease Control and Prevention, World Cancer Day, February 3, 2015.
2. Department of Health and Human Services Centers for Disease Control and Prevention, United States Cancer Statistics, Technical Notes 2007.
3. American Cancer Society, Cancer Facts & Figures 2016, Atlanta, Georgia, American Cancer Society, pp. 1– 63, 2016.
4. Simes RJ. Treatment selection for cancer patients: application of statistical decision theory to the treatment of advanced ovarian cancer. J Chronic Dis, 38:171-86, 1985.
5. Maclin PS, Dempsey J, Brooks J, et al. Using neural networks to diagnose cancer J Med Syst, 15:11-9, 1991.
6. Cicchetti DV. Neural networks and diagnosis in the clinical laboratory: state of the art. Clin Chem, 38:9-10, 1992.
7. Petricoin EF, Liotta LA. SELDI-TOF-based serum proteomic pattern diagnostics for early detection of cancer. Curr Opin Biotechnol, 15:24-30, 2004.
8. Bocchi L, Coppini G, Nori J, Valli G. Detection of single and clustered micro calcifications in mammograms using fractals models and neural networks. Med Eng Phys, 26:303-12, 2004.

9. Zhou X, Liu KY, Wong ST. Cancer classification and prediction using logistic regression with Bayesian gene selection. J Biomed Inform, 37:249-59, 2004.
10. Dettling M. Bag Boosting for tumor classification with gene expression data. Bioinformatics, 20:3583-93, 2004.
11. Wang JX, Zhang B, Yu JK, et al. Application of serum protein finger printing coupled with artificial neural network model in diagnosis of hepatocellular carcinoma. Chin Med J (Engl), 118:1278-84, 2005.
12. McCarthy JF, Marx KA, Hoffman PE, et al. Applications of machine learning and high-dimensional visualization in cancer detection, diagnosis, and management. Ann N Y Acad Sci, 1020:239-62, 2004.
13. L. Breiman. Bagging predictors. Machine Learning, 24:123–140, 1996.
14. P. Buhlmann and B. Yu. Analyzing bagging, The Annals of Statistics, 30:927–961, 2002.
15. A. Buja and W. Stuetzle. Observations on bagging. Statistica Sinica, 16:323–352, 2006.
16. G. Biau, F. Cerou, and A. Guyader. On the rate of convergence of the bagged nearest neighbor estimate. Journal of Machine Learning Research, 11:687–712, 2010.