CERVICAL CANCER DETECTION USING MACHINE LEARNING TECHNIQUES

Abstract

Cervical cancer is frequently a deadly disease, common in females. However, early 
diagnosis of cervical cancer can reduce the mortality rate and other associated 
complications. Cervical cancer risk factors can aid the early diagnosis. For better 
diagnosis accuracy, we proposed a study for early diagnosis of cervical cancer using 
reduced risk feature set and three ensemble-based classification techniques, i.e., 
extreme Gradient Boosting (XGBoost), AdaBoost, and Random Forest (RF) along 
with Firefly algorithm for optimization. Synthetic Minority Oversampling Technique 
(SMOTE) data sampling technique was used to alleviate the data imbalance problem. 
Cervical cancer Risk Factors data set, containing 32 risks factor and four targets 
(Hinselmann, Schiller, Cytology, and Biopsy), is used in the study. The four targets 
are the widely used diagnosis test for cervical cancer. The effectiveness of the 
proposed study is evaluated in terms of accuracy, sensitivity, specificity, positive 
predictive accuracy (PPA), and negative predictive accuracy (NPA). Moreover, 
Firefly features selection technique was used to achieve better results with the reduced 
number of features. Experimental results reveal the significance of the proposed 
model and achieved the highest outcome for Hinselmann test when compared with 
other three diagnostic tests. Furthermore, the reduction in the number of features has 
enhanced the outcomes. Additionally, the performance of the proposed models is 
noticeable in terms of accuracy when compared with other benchmark studies for 
cervical cancer diagnosis using reduced risk factors data set.