Skin Cancer Prediction Using Three Major Classification Algorithms

Abstract

Skin Cancer Classification is a major public health concern, and correctly classifying skin lesions as benign or malignant is critical for accurate diagnosis and treatment. This text provides a comprehensive guide to building and evaluating models to accurately classify skin lesions as benign or malignant. It includes an introduction, overview, data collection and preparation, setting up the environment, data preprocessing, splitting the data into train and test sets, scaling the data, models (SVM, RF and KNN), visualization and analysis, and conclusion. The project used three classification algorithms: Support vector machines, Random Forest, and K-Nearest Neighbors (KNN) to create models for skin cancer classification. Results showed that Random Forest had the highest accuracy of 98%, followed by SVM with 94% and KNN with 74% accuracy. This initiative advances the field of skin cancer diagnostics by making use of classification algorithms