We are developing a method to diagnose diabetes with the help of machine learning algorithms and tools. The proposed method utilizes the power of machine learning to create a model that can predict diabetes based on the health behaviors of the patient. The model makes use of the relationship between a healthy lifestyle and diabetes, a healthy lifestyle helps prevent diabetes and the opposite holds true. This relationship has been supported by many studies. Our goal is to build a machine learning model with adequate accuracy to predict diabetes which will help significantly in easing and speed up the diagnosing procedure of diabetes. We used relatively new machine learning algorithms to build the model namely: the top three Gradient Boosting algorithms XGBoost, LightGBM, CatBoost, and Artificial Neural Networks.
The dataset that we will use is been collected from 6 waves of the National Health and Nutrition Examination Survey: 2007-2008, 2009-2010, 2011-2012, 2013-2014, 2015-2016, and 2017-2018.
https://wwwn.cdc.gov/nchs/nhanes/Default.aspx
The combined dataset is consisting of 30 features and 55939 observations and all values of the dataset have been coded in numerical form.