Machine Learning Diabetes Diagnosis Literature Review

This paper presents a systematic literature review on the use of machine learning in diagnosing Diabetes Mellitus (DM). The study examines the application of machine learning algorithms and datasets in diabetes research. The findings highlight the effectiveness of Random Forest and the prevalence of the PIMA Indian dataset in this field. Early detection of diabetes is crucial for effective management and prevention of complications. However, challenges such as limited healthcare access and undiagnosed cases exist. The analysis reveals challenges related to dataset quality, sensitivity-specificity trade-offs, outliers, and missing data. To overcome these challenges, future research should expand the training dataset, incorporate additional parameters, and address outlier handling techniques. Feature selection methods and careful consideration of sensitivity-specificity trade-offs are also recommended. Despite these challenges, machine learning has the potential to improve diabetes diagnosis and enhance medical care. This study provides valuable insights for future advancements in machine learning-based diabetes diagnosis.