Predicting Skin Cancer using Machine Learning

Skin cancer affects an increasing number of people which is the main motivation for my research. The goal is to develop a model that is accurate enough for real world application and that could be distributed to more rural areas which have decreased access to medical technology.

The motivation for this research project was to create an accurate model that could be used to identify the presence of different types of skin cancers. The latter objective is to create an application that could be easily used in rural areas, as they do not have access to modern medical technology. The approach began with loading the dataset into a readable format. I used the load_dataset import from Hugging Face [4] to complete this task. The models that were used for this research project were all from Hugging Face [4]. The highest accuracy was 0.8633 and the lowest accuracy was 0.6647. The models were all run on a colab notebook. The meaning of these results is that on the best model the image was predicted correctly 86% of the time. The conclusion of this research is that more research must be conducted and more models must be used in order to achieve an accuracy that would warrant the model to be used in the real world.

Explore More!

Research Paper