An Efficient Technique for the Better Recognition of Oral Cancer using Support Vector Machine

Accuracy is among the most important factors in a disease diagnosis. It is essential to select the characteristics that you find most pertinent for the highest accuracy. This study aims to more accurately predict the presence of the primary stage of squamous cell carcinoma using fewer indicators. Stages of oral cancer were first demonstrated to be predicted by 25 features. The variety of features that are obtained from various patient records is indirectly decreased in this study through the combination of the unified medical system using hybrid features selection techniques to identify the characteristics that are most useful for the identification of oral cancer. Hybrid feature selection has been used to condense 25 qualities into 14 features. The diagnosis of patients with oral cancer is then predicted using four classifiers: Updatable Naive Bayes, Multilayer Perceptrons, K-Nearest Neighbor, and Support Vector Machines. Also, the data show that, after adding feature development decisions with SMOTE during the preprocessing phases, the support vector machine’s performance surpasses other machine learning techniques.