“COMPARISON OF MACHINE LEARNING TECHNIQUES FOR ACCURATE DIABETES RISK PREDICTION”

Authors:

Abdul Aamir Khan, Dr. Balkrishna Sharma

Page No: 472-480

Abstract:

Despite being one of the most common diseases in the world, diabetes can be cured and prevented if identified early. Ultimately, we want to find a method that reliably solves categorization problems with accurate results. In order to pre-process the data and extract useful characteristics, feature engineering approaches were used. A variety of machine learning algorithms, including Decision Trees, Random Forests, and Support Vector Machines, were employed to effectively develop models for predicting diabetes risk assessments. Another result of looking into disease influence measurements was finding out what causes diabetes primarily. Using feature importance and correlation coefficient analysis, this study seeks to illuminate the relative significance of age, BMI, family history, and glucose levels as risk factors. When it comes to applications that require a balance between recall and precision, Naive Bayes has proven to be adaptable across all metrics, while Support Vector Machines (SVM) has proven to be the most reliable model with the best accuracy.

Description:

.

Volume & Issue

Volume-13,ISSUE-12

Keywords

Keywords: Diabetes, Machine learning, Accuracy, Precision, Prediction