Hybrid intelligent system-based rough set and ensemble classifier for breast cancer diagnosis

The effectiveness of classification and recognition
systems has improved in a great deal to help medical
experts in diagnosing diseases. Breast cancer is becoming a
leading cause of death among women in the whole world;
meanwhile, it is confirmed that the early detection and
accurate diagnosis of this disease can ensure a long survival
of the patients. This paper presents a hybrid intelligent
system for recognition of breast cancer tumors. The
proposed system includes two main modules: the feature
extraction module and the predictor module. In the feature
extraction module, rough set theory is used to preprocess
the attributes on condition that the important information is
not lost, deletes redundant attributes and conflicting objects
from decision table. In the predictor module, a combined
classifier is proposed based on K-nearest neighbor classifier.
Experiments have been conducted on a widely used
Wisconsin breast cancer dataset taken from University of
California Irvine. Experimental results show that the proposed
hybrid system can improve the rate of correct
diagnosis of cases. The proposed combined classifier with
rough set-based feature selection achieves 99.41 % classification
accuracy and uses only 4 features which is the
best shown to date. Different performance metrics are used
to show the effectiveness of the proposed hybrid system.
With these results, the proposed method is very promising
compared to the previously reported results and can be
used confidently for other breast cancer diagnosis
problems.