Tehran University of Medical Sciences

Science Communicator Platform

Stay connected! Follow us on X network (Twitter):
Share this content! On (X network) By
Prediction Breast Cancer Risk: Performance Analysis Data Mining Techniques Publisher



Sohrabi S1 ; Atashi A2
Authors
Show Affiliations
Authors Affiliations
  1. 1. Department of Medical Informatics, Shahid Beheshti University of Medical Sciences, Tehran, Iran
  2. 2. Department of E-Health, Virtual School, Tehran University of Medical Sciences, Medical Informatics Department, Breast Cancer Research Center, Motamed Cancer Institute (ACECR), Tehran, Iran

Source: Frontiers in Health Informatics Published:2021


Abstract

Introduction: Early detection breast cancer Causes it most curable cancer in among other types of cancer, early detection and accurate examination for breast cancer ensures an extended survival rate of the patients. Risk factors are an important parameter in breast cancer has an important effect on breast cancer. Data mining techniques have a growing reputation in the medical field because of high predictive capability and useful classification. These methods can help practitioners to develop tools that allow detecting the early stages of breast cancer. Material and Methods: The database used in this paper is provided by Motamed Cancer Institute, ACECR Tehran, Iran. It contains of 7834 records of breast cancer patients clinical and risk factors data. There were 4008 patients (52.4%) with breast cancers (malignant) and the remaining 3617 patients (47.6%) without breast cancers (benign). Support vector machine, multi-layer perceptron, decision tree, K nearest neighbor, random forest, naive Bayesian models were developed using 20 fields (risk factor) of the database because database feature was restrictions. Used 10-fold crossover for models evaluate. Ultimately, the comparison of the models was made based on sensitivity, specificity and accuracy indicators. Results: Naive Bayesian and artificial neural network are better models for the prediction of breast cancer risks. Naive Bayesian had accuracy of 93%, specificity of 93.32%, sensitivity of 95056%, ROC of 0.95 and artificial neural network had accuracy of 93.23%, specificity of 91.98%, sensitivity of 92.69%, and ROC of 0.8. Conclusion: Strangely the different artificial intelligent calculations utilized in this examination yielded close precision subsequently these techniques could be utilized as option prescient instruments in the bosom malignancy risk considers. The significant prognostic components affecting risk pace of bosom disease distinguished in this investigation, which were approved by risk, are helpful and could be converted into choice help devices in the clinical area. © 2021, Published by Frontiers in Health Informatics.
Related Docs
1. A Decision Support System for Mammography Reports Interpretation, Health Information Science and Systems (2020)
2. Prediction of Breast Cancer Using Machine Learning Approaches, Journal of Biomedical Physics and Engineering (2022)
Experts (# of related papers)