Paper Contents
Abstract
Skin disease recognition and monitoring present significant challenges in the medical industry. The increasing levels of environmental pollution and unhealthy dietary habits have contributed to a sharp rise in skin- related issues. Beyond health concerns, poor skin conditions can negatively impact an individual's confidence. Early detection through regular and accurate skin monitoring is essential to prevent the progression of serious skin diseases. Machine learning techniques offer promising solutions for developing robust systems capable of classifying various skin conditions.A crucial step in this process is distinguishing between skin and non-skin regions to ensure accurate diagnosis. In this study, we implemented and analyzed five different machine learning algorithmsRandom Forest,Nave Bayes, Logistic Regression, Kernel SVM, and Convolutional Neural Networks (CNN)on a skin disease dataset to predict disease categories. A comparative evaluation was conducted based on confusion matrix parameters and trainingaccuracy, with results visualized using graphical representations. Among these, CNN demonstrated the highest accuracy and proved to be the most effective for skin disease classification. Additionally, the proposed system not only predicts the type of skin disease but also provides precautionary measures to assist patients in managing and preventing further complications.
Copyright
Copyright © 2025 E Bharghav Rao. This is an open access article distributed under the Creative Commons Attribution License.
