Neutrosophic Statistical Analysis on Healthcare Data for Oral Cancer Detection

Sakshi Taaresh Khanna^1,*, Sunil Kumar Khatri², Neeraj Kumar Sharma³

¹Amity Institute of Information Technology, Amity University, Uttar Pradesh, Noida-201313, India

²Amity University, Uttar Pradesh, Noida-201313, India

³Department of Computer Science, Ram Lal Anand College, Benito Juarez Marg, New Delhi-110021, India

Emails: sakshitaareshkhanna@gmail.com; skkhatri@amity.edu; neerajksharma100@gmail.com

Abstract

Oral cancer is presently a growing health concern at the global level, with intense incidences of lifestyle factors. The increasing mortality rates of the diseased shall be controlled with effective early detection mechanisms. However, the traditional statistical approaches in practice fail to deliver in making a precise diagnosis of this cancer due to the intricate and interdependent prevalence of symptoms. This research work provides a solution approach using the potency of neutrosophic statistics in developing neutrosophic-integrated models of random forests and decision trees. Neutrosophic representation of data considering the indeterminacy, values of truth, and falsity facilitates healthcare experts in handling the conflicting patient data. The proposed random forest decision model with neutrosophic logic identifies the significant features, and the neutrosophic decision tree classifier predicts the stages of cancer. The findings are compared with conventional modelling of random forest and decision trees, and it demonstrates the efficiency and precision of neutrosophic statistical analysis in predicting oral cancer. This proposed neutrosophic decision framework will assist and support the medical practitioners and research experts in gaining more insights and deeper comprehension of the cancer progression and suggesting suitable treatment plans to minimize the morbidity rate.

Keywords: Oral cancer; Neutrosophic logic; Neutrosophic Random Forest Algorithm; Neutrosophic decision trees; Prediction