Utilising Machine Learning Techniques to Evaluate the Survival Probability of Head and Neck Cancer Patients Based on Their Smoking Habits and Medical History

Malignancies are prevalent worldwide in the contemporary era, encompassing conditions affecting the oral cavity, throat, and vocal cords. Smoking and tobacco use can be regarded as a primary factor contributing to the development of malignant malignancies in the patients. This research paper presents a novel method for examining the relationship between lifestyle patterns and cancers using key cancer indicators such as the TNM cancer staging system and HPV cancer screening. Therefore, it is necessary to develop an algorithm that can determine the anticipated fatality rates among patients with head and neck cancer. These algorithms help physicians identify the key factors that contribute to patients' prognosis, enabling them to deliver appropriate treatment. The authors of this work utilised all the models and ensured that all of them were fine-tuned including hyperparameters. The gradient boosting algorithm achieved the highest level of precision, with a reliability of up to 98.8 %. According to this classifier, smoking and HPV positive were shown to be the major risk factors, with rates of approximately 4 % and 2 %, respectively, with a significance level of 5 %. To validate the performance of classifiers, the gradient boosting approach achieved a maximum value of 0.99, which signifies exceptional performance. Within this framework, this study will provide significant benefits to healthcare practitioners who are involved in the care of cancer patients who are experiencing challenges with their life expectancy. It will enhance their ability to predict the mortality rates of these patients more effectively, hence enabling them to provide tailored and more precise treatment strategies.