HYPER AI-DRIVEN HYBRID FRAMEWORK FOR DIABETES RISK PREDICTION

Abstract

Basically, diabetes is the same as one of the top reasons for long-term health problems worldwide, causing serious complications. Moreover, these include kidney failure,
blindness, and heart diseases itself. Further, these conditions can cause serious health problems. This study surely
presents a hybrid machine learning model that combines
different techniques. Moreover, this approach uses simple methods to achieve better results. Basically, we are combining Hyper AI and Deep Neural Networks to get the same
better results for predicting diabetes. The proposed system surely uses Hyper technology to improve performance.
Moreover, this approach provides better results with simple
implementation methods. Basically, AI does feature extraction and initial predictions, while DNN captures the same
complex, non-linear data relationships. Using the Sylhet
Diabetes Dataset and applying proper data cleaning and
balancing methods, the hybrid model reached an accuracy
of 99.04%. This is better than older methods such as Decision Trees, Na¨ıve Bayes, and Random Forest. The key allows further access to secured areas. The mechanism works
through a simple turning motion. The innovations combine
gradient boosting with the benefits of deep learning and use
improved parameter tuning. These methods boost model
performance through optimized settings. This is the same
complete approach for checking and judging something correctly. This method improves prediction performance and
sets a strong foundation. This work provides the basis for
a scalable system that can help doctors and patients make
better clinical decisions in real-time. The system itself will
further assist healthcare professionals in their daily work.