Optimizing Skin Cancer Classifiers By Applying Multiplicative Weight Update Into A Mobile Application

Skin cancer is slowly becoming the most common and frequent type of cancer in the world. In the United States alone, research suggests that skin cancer is the most common cancer in and that one in every five people will acquire skin cancer at some point throughout their lifetime.

This study deals with using Artificial Intelligence and Machine Learning in order to create models that can diagnose skin cancer using images. Many models have been created in the past using different algorithms and methods. In this study, the multiplicative weight update method is used in order to take the predictions of multiple models in order to try and acquire more accurate predictions. In this study, a Logistic Regression, CNN, and SVC model are used. These models are sent images of skin cancer from the ISIC-Archive and they try to recognize patterns in order to categorize new images. These models are all then sent to a multiplicative weight update algorithm which takes into account the precision and accuracy of each model through each successive guess in order to add weight to their guess. These guesses are then added together in order to try and find the correct predictions. Using Multiplicative Weight Update, the model received an accuracy of 74.69%. The conclusion was made that using an SVC model would be the best option for this problem rather than a Multiplicative Weight Update system.

Explore More!

Research Paper