Enhanced Diagnosis of Gestational Diabetes through Machine Learning and Logistic Regression

A persistent illness Diabetes might become an international health emergency. Diabetes mellitus, also known as diabetes, is typified by high blood sugar levels. Being among the worst diseases in the world, it is also spreading more and more. Diabetes can be diagnosed by a variety of traditional techniques as well as chemical and physical testing. Techniques used in data science may benefit other scientific disciplines since they throw new insight on pressing problems. Machine learning is a contemporary area of data science that focuses on how computers learn from experience. This paper develops and assesses a machine learning model to predict gestational diabetes using a large set of data of physiological markers and health features This study uses 768 examples to create and test a binary classification model to predict gestational diabetes utilizing characteristics including pregnancy, blood pressure, insulin levels, BMI, age, and family history. Class '1' diabetes has 0.72 accuracy, while class '0' has 0.79. However, diabetic recall is substantially poorer (0.55) than non-diabetic (0.89), indicating many missed positives. F1-scores (0.83 for class '0' and 0.63 for class '1') indicate the model's diabetes detection shortcomings and the need for prediction improvement.