Optimized Diabetes Prediction using Logistic Regression with SGD and MBGD

Optimization algorithms are essential for machine learning models to enhance prediction accuracy. There is a surge in the number of people suffering from diabetes in this present day, it is possible to forecast the possibility of diabetes via logistic regression use of two prominent optimization methods of SGD and MBGD. The proposed method of diabetes prediction consists of processing of the data, selecting the relevant features and configuration of hyper parameters for the optimization of the model. The primary objective is to compare the effectiveness of logistic regression for both SGD and MBGD, to discriminate diabetics and non-diabetics across a wide population whose demographic and clinical particulars are known. The experimental validation of MBGD in terms of accuracy outperforms better than the SGD. Finally, this model of diabetes prediction, assessed according to accuracy, sensitivity and F1 score achievements, is effective in detection since it has a high level of overall accuracy, with low false positive rates and high levels of efficient true positive detections. The proposed optimized logistic regression model provides healthcare professionals with a powerful tool for timely intervention and personalized patient care. The results highlight the model and its promise for integration into clinical practice, leading to more efficient and accurate diabetes diagnosis and treatment. This analysis advances the field by improving predictive accuracy and computational efficiency in diabetes prediction.