Using Supervised Machine Learning to Forecast Cancer Death Cases

Abstract

Cancer kills a disproportionate number of people in India compared to other countries. Using supervised machine learning techniques, this study aims to forecast cancer death rates in India. Using data from the Global Burden of Disease Study, we can see the cancer death rates in India from 1990 to 2017 broken down by gender, age group, and location. Following the completion of data pretreatment procedures such as feature engineering and missing value imputation, we use three separate supervised learning algorithms: random forest regression, decision tree regression, and linear regression. We compare the three models' performance using a number of metrics and find that the random forest regression model is the best. In order to aid the health department in their efforts, the study aims to use the best model to offer a long-term forecast of cancer mortality in India. Policymakers and healthcare professionals in India may use our study to influence their efforts to enhance cancer care and decrease cancer incidence. Subjects—Cancer in India, Cancer Deaths, Prediction, ML with Supervision, Linear and Decision Tree Regression, RF and Polynomial Regression, and Measures for Success