Can Breast Cancer Survival be predicted by Risk Factors? Machine Learning Models

Montazeri, Mitra and Montazeri, Mohadeseh and Montazeri, Mahdieh and Bahrampour, Abbas (2015) Can Breast Cancer Survival be predicted by Risk Factors? Machine Learning Models. In: 10th International Breast Cancer Congress, 25-27 February 215, Tehran, Iran.

[img]
Preview
Text
International Breast Cancer Congress.pdf - Published Version

Download (13MB) | Preview
Official URL: http://ibcc.sbmu.ac.ir/

Abstract

Breast cancer is a kind of cancer with high mortality among women. With early diagnosis of breast cancer (up to five years after cell division) survival will increase from 56% to more than 86%. Therefore, an accurate and reliable system is necessary for early diagnosis of benign or malignant tumors. Automatic classification systems as a diagnostic tool can reduce the workload of doctors. Intelligent methods to predict Breast cancer survival which are used in this study consist of Naïve Bayes, Trees Random Forest, 1NN, AdaBoost, SVM, RBF Network and Multilayer Perceptron. In this study 900 patient records are used. These records have been registered at Cancer Registry Organization of Kerman Province, in Iran. For evaluate the proposed models, K-fold cross validation is used. Seven models of machine learning are compared base on specificity, sensitivity and accuracy. The accuracy of the seven models are .95%, .96%, .91%, .94%, .94%, .95% and .95% respectively. Our result showed that trees Random Forest model was the best model with the highest level of accuracy. Therefore, Trees Random Forest model is recommended to Breast cancer survival.

Item Type: Conference or Workshop Item (Paper)
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
T Technology > T Technology (General)
Depositing User: mitra montazeri montazeri
Date Deposited: 01 Dec 2015 09:12
Last Modified: 20 Feb 2016 07:50
URI: http://eprints.kmu.ac.ir/id/eprint/21520

Actions (login required)

View Item View Item