A study of US expenditures on cancer treatment with data analysis and machine learning

Cancer is the second leading cause of death around the world, causing cancer cost to be an important social issue in the United States. News reports show that American cancer patients spent more than $21 billion on their care in 2019. (US News, 2021) In this research, data analysis has been done based on the national expenditure on cancer treatment from 2010 to 2020 through the use of Python language and available third party libraries. Also, a machine learning classification model has been trained, developed and tested to help predict the cost of cancer treatment in the next few years. Among four different machine learning regression algorithms that are applied (i.e linear regression, lasso regression, random forest regression, and gradient boosting regression), gradient boosting regression is the best fit for the model, aiming to produce the most accurate prediction to inform people and government officials.