Predicting Melanoma Tumor Size through Machine Learning Approaches
Dhruv Chadha, Nikita Jain, Vedika Gupta · 2021
Melanoma is a kind of skin cancer that arises from melanocytes, pigment-producing cells. Melanoma is more harmful than other cancers because it can spread to other organs more quickly if not treated early. As a result, it is important to detect melanoma early. Existing research has focused on predicting whether or not a tumor is present, but no such methodology has been used to predict tumor size. This chapter focuses on measuring the tumor&s;s size based on the various features provided. The size of a tumor has been predicted using a number of machine learning methods. RMSE, MAE, R2 score, and MBE were used to assess the prediction&s;s accuracy. Based on the results obtained, combined model of XGBoost, LightGBM, Extra Tree Regressor, Catboost, and Bagging Regressor outperforms all approaches.