ENHANCED DERMATOSCOPIC SKIN LESION CLASSIFICATION USING MACHINE LEARNING TECHNIQUES

Authors

  • Khutaija Abid Assistant Professor, Department of IT, Lords Institute of Engineering and Technology, Hyderabad Author
  • Sibghatullah, Mustafa Ali, Shaik Obaid Rahman B.E. Student, Department of IT, Lords Institute of Engineering and Technology,Hyderabad Author

Abstract

Malignant melanoma poses significant impact on public health. Extensive research has focused on
distinguishing between benign and malignant skin lesions through dermatoscopic image analysis. Our study
emphasizes the classification aspect, using the MNIST HAM 10000 dataset. Initially, we addressed the challenge of
imbalanced data by applying Synthetic Minority Oversampling Technique (SMOTE) to balance the dataset,
significantly improving accuracy across various machine learning algorithms. Among these algorithms—Decision
Tree (using Gini index and Entropy), Naïve Bayes, XGBoost, Random Forest, Logistic Regression, and Support
Vector Machine (specifically with Polynomial kernel)—we found that Support Vector Machine with Polynomial
kernel achieved the highest accuracy of 96.825%. While XGBoost, being a Gradient Boosting algorithm, showed
varying results, we verified its accuracy using k-fold cross validation (k=10), achieving 95.984%. Ultimately, our
findings highlight Support Vector Machine with Polynomial kernel as the most effective for this classification
task[1]

Downloads

Published

2024-09-30

Issue

Vol. 14 No. 7 (2024): September

Section

Articles

How to Cite

ENHANCED DERMATOSCOPIC SKIN LESION CLASSIFICATION USING MACHINE LEARNING TECHNIQUES. (2024). INTERNATIONAL JOURNAL OF MANAGEMENT RESEARCH AND REVIEW, 14(7), 32-37. https://ijmrr.com/index.php/ijmrr/article/view/269

Most read articles by the same author(s)