International Journal of Advanced Innovative Technology in Engineering (IJAITE)

---

Detection and Prediction of Skin Cancer based on Machine Learning

Nilesh S. Jadhao, Dr. Ms. S. W. Ahmed

Abstract :

Statistical evidence suggests that skin related deaths were most in number compared to those resulting from other types of skin cancer. However, detection of skin cancer in its early stages is extremely challenging to doctors since as they say cancer is skin-deep. Wrong diagnosis of the disease due to inaccurate methods has prompted study in this field for a while since malignant melanoma implies asymmetrical and irregular borders, notched edges and color variations. Hence, it is extremely vital to analyse each characteristic of the lesion like shape, colour and texture early detection and prevention. In this study, effective detection of skin cancer detection is proposed. Four features are chosen that are trained and tested by using various classification techniques like K-Nearest Neighbor, Decision Tree, Naive Bayes, Random Forest, and XGBoost have been done. The methodology actually has a good result for the various classifiers but still can be improved. The result discusses that XGBoost is a better classifier than K-Nearest Neighbor, Decision Tree, Naive Bayes, Random Forest has the highest accuracy of 68.75%. The results achieved are relatively good when compared. The future work will focus on improving the number of features selected and then classified to improve the accuracy.

Keywords :

Full Text :

Download PDF

DOI :

Cite this paper :

References :

---