Breast Cancer Prediction Using Machine Learning

Parinitha R, Sangeetha N, Pooja, Punith M; Students of ISE Dept

Abstract: Results and Discussion:
Machine learning is used in detection of cancerous i. Data Exploration : The distributions of the mean,
cells. Cancerous cells are classified as Benign (B) or standard error and worst average of the 10
Malignant (M). There are many algorithms for features extracted from the FNA slides show that
classification and prediction of breast cancer. In this compactness, concavity, symmetry etc.
project, Support Vector Machine (SVM) on the ii. Correlation: iii. Count of Benign(B)
Wisconsin Breast Cancer dataset is used. The dataset is and Malignant(M):
also trained with the other algorithms: KNN, Naives
Bayes and CART.

Introduction:
Breast Cancer is type of cancer that occurs mostly in
females leading to death. These deaths can be iv. Performance Comparison:
reduced by early detection of cancerous cells.
GNB, KNN & CART : 92% mean accuracy
Cancerous cells are detected by tests like MRI, Biopsy.
SVM : Low performance, however it can be
Improved by standardizing the input.
The performance of this algorithm is expected to be
Materials & Methodology: High given the symptoms for breast cancer should
Python Software for coding. Breast cancer data is from Exhibit certain clear patterns.
UCI Depository. And methodologies are ML techniques
such as SVM, KNN, Decision Trees & Naives Bayes. Conclusion and Future Scope:
KNN & Naïve Bayes give best results. SVM is strong
Dataset: technique for predictive analysis. We conclude that
SVM using Gaussian Kernel is the most suitable
Dataset contains 357 cases of Benign breast cancer techniques for prediction of breast cancer. Fine tuning
and 212 cases of Malignant breast cancer. The dataset of parameters used in algorithms can result in better
contains Id No, Diagnosis result then, mean, SD. accuracy. Furthermore, this can also be implemented
Features like radius, texture, perimeter etc. are on a cloud platform for easier usage.
obtained from digitized image of FNA Biopsy of
tumor.
References:
Patient [1] UCI depository - http://archive.ics.uci.edu/ml/ [2]
Maglogiannis, I., Zafiropoulos, E., & Anagnostopoulos
- An intelligent system for automated breast cancer
Data Formation diagnosis and prognosis using SVM based classifiers,
Applied intelligence journal, Volume 30,Issue 1,
February 2009 [3] S.Kharya, D.Dubey, S.Soni -
Training Data Testing Data Predictive Machine Learning Techniques for Breast
Cancer Detection, IJCSIT) International Journal of
Computer Science and Information Technologies, Vol.
4 (6), 2013, 1023-1028 [4] Mandeep Rana, Pooja
i. Support Vector Machine Chandorkar, Alishiba Dsouza, Nikahat Kazi – Breast
ii. CART cancer Diagnosis and Recurrence prediction using
iii. KNN Machine learning Techniques,IJRET- International
Journal of Research in Engineering and Technology,
iv. Naïve Bayes April 2015 [5]National Breast Cancer Foundation Inc.,
http://www.nationalbreastcancer.org/about-breast-
cancer.
Result Comparison