Prognostication of Diabetes Diagnosis Based on Different Machine Learning Classification Algorithms

Abstract

Diabetic disease is one of the most deadly and infectious illnesses that induce blood glucose (sugar) levels to rise. If this disease is left unchecked and undiagnosed, it can lead to a slew of consequences. Diabetes is the creator of various diseases like diabetic retinopathy, heart problems, etc. The time-consuming identification procedure leads to a patient's initial consultation to a clinical facility and consultation with a practitioner. Data science techniques always have opportunity to aid other scientific areas by avoiding existing concerns on traditional issues. Machine learning is a recent field of AI which looks at how machines learn via their experiences. The emergence of ML strategies on the other hand, addresses this crucial issue. If accurate earlier detection is achievable, the health risk and intensity of disease may be greatly decreased. Owing to the small range of labelled records and the inclusion of data points and missing values in diabetic databases, stable and reliable diagnosis prognostication is extremely difficult. The aim of this research is to build a model that can accurately predict the risk of developing diabetes in individuals. Exploratory data analysis, data pre - processing, feature importance, feature engineering, and different Machine Learning classifiers are all part of a plan that we introduce for diabetes to diagnose.  For diabetes prognostication, our paradigm outshines the various strategies mentioned in the study. It can also provide better results from the same data source, leading to better diagnostic of prediction performance.