Predicting Diabetes Disease Occurrence Using Logistic Regression: An Early Detection Approach

Abstract

 Diabetes disease is prevalent worldwide, and predicting its progression is crucial. Several model have been
proposed to predict such disease. Those models only determine the disease label, leaving the likelihood of developing the disease
unclear. Proposing a model for predicting the progression of disease becomes essential. Therefore, this article proposes a logistic
regression model to anticipate the likelihood of Diabetes syndrome incidence. The model exploit capabilities of logistic regression
by using sigmoid function. The model's performance was evaluated using the Pima Indians Diabetes dataset and demonstrated
high accuracy, sensitivity, and specificity. The prediction accuracy rate was 77.6%, with a sensitivity of 72.4%, specificity of
79.6%, Type I Error of 27.6%, and Type II Error of 20.4%. Furthermore, the model indicates the feasibility of using laboratory
tests, such as Pregnancies, Glucose, Blood Pressure, BMI, and DiabetesPedigreeFunction, to predict disease progress. The
proposed model can aid patients and physicians in understanding the disease's progression and implementing timely interventions