• Version
• Download 13
• File Size 504.41 KB
• File Count 1
• Create Date 5 May 2025
• Last Updated 5 May 2025

AI in Diabetes Prediction and Management

ABHINAV GOEL
Department of AI-ML ADGIPS - Delhi New Delhi, INDIA

rabhigoyal31@gmail.com

AYUSH GUPTA
Department of AI-ML ADGIPS - Delhi New Delhi, INDIA

reach2ayushgupta@gmail.com

HIMANSHU RAVI
Department of Al-ML ADGIPS - Delhi New Delhi, INDIA

himanshukashyap9122@gmail.com

SHUCHI SHARMA
Department of Al-ML ADGIPS - Delhi New Delhi, INDIA

303shuchi@gmail.com

Abstract:
The increasing awareness of health and wellness has prompted extensive research on the impact of diet, physical activity, and lifestyle choices on overall well-being. This study aims to analyze health-related trends using a survey-based dataset, collected through a structured Google Form. The survey included demographic details (age, gender, occupation, height, weight) and lifestyle-related factors such as dietary habits, exercise frequency, and daily routines. The responses were visualized using pie charts to identify trends among different groups.

To extract meaningful insights, the collected data underwent preprocessing—removal of inconsistencies, handling of missing values, and transformation of categorical responses into numerical values for machine learning analysis. Various machine learning models, including Logistic Regression, Decision Trees, Random Forest, Support Vector Machine (SVM), and Neural Networks (MLP), were trained to classify health risk levels based on survey responses. Among these, the Random Forest model achieved the highest accuracy (89.4%), outperforming other models in terms of precision (0.88), recall (0.87), and F1-score (0.87). The Neural Network (MLP) also exhibited strong performance with 87.2% accuracy and a high AUC-ROC score of 0.90, indicating its reliability for predictive health analytics.

This research emphasizes the importance of data-driven health assessments and the effectiveness of machine learning in analyzing lifestyle factors. Future work can extend this study by incorporating real-time health tracking, larger datasets, and deep learning models for enhanced predictive accuracy.