Breast cancer recurrence prediction using machine learning

Kaustubh Chakradeo, Sanyog Vyawahare, Pranav Pawar

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

12 Scopus citations

Abstract

The most common cancer among women is breast cancer. Around 12% of women are affected by it all over the world. Recurrent breast cancer is a term used for breast cancer which returns even after a successful treatment. This research aims to use Machine learning to detect and predict the recurrence of breast cancer; and compare all the models by using different metrics like accuracy, precision, etc. The models built can help predict the recurrence of breast cancer effectively. All the models are built using the Wisconsin Prognostic Breast Cancer Dataset(WPBC). The models built are Multiple Linear Regression, Support Vector Machine, which was build by using RBF Kernel and Leave-One-Out(K-fold Cross-Validation) and Decision Tree using metrics like Gini Index, Entropy and Information Gain. Support Vector Machine and K-fold Cross-Validation gave the best results for recurrence and non-recurrence predictions.

Original languageEnglish
Title of host publication2019 IEEE Conference on Information and Communication Technology, CICT 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728153988
DOIs
StatePublished - Dec 2019
Event2019 IEEE Conference on Information and Communication Technology, CICT 2019 - Allahabad, India
Duration: 6 Dec 20198 Dec 2019

Publication series

Name2019 IEEE Conference on Information and Communication Technology, CICT 2019

Conference

Conference2019 IEEE Conference on Information and Communication Technology, CICT 2019
Country/TerritoryIndia
CityAllahabad
Period6/12/198/12/19

Bibliographical note

Publisher Copyright:
© 2019 IEEE.

Keywords

  • Breast Cancer
  • Decision Tree
  • Prognosis
  • Recurrence
  • Regression
  • SVM

Fingerprint

Dive into the research topics of 'Breast cancer recurrence prediction using machine learning'. Together they form a unique fingerprint.

Cite this