Accurate Prediction of Diabetes Risk Using a Multi-Scale Temporal Long Short-Term Memory Network

Description

AD-STLSTM is an advanced deep learning framework designed for accurate prediction of diabetes risk using patient-level clinical and symptomatic data. The model leverages a Multi-Scale Temporal Long Short-Term Memory (ST-LSTM) architecture to effectively capture complex, nonlinear relationships among diabetes-related symptoms and demographic factors.

By processing input features at multiple temporal and representational scales, AD-STLSTM can model both short-term symptom interactions and long-range dependencies that contribute to diabetes onset. This multi-scale learning strategy enhances feature discrimination and improves robustness against noisy or overlapping clinical indicators.

The proposed AD-STLSTM framework enables early and reliable identification of diabetes risk, providing an AI-driven decision support tool for preventive screening and clinical diagnosis.

Input

1. Age
2. Gender
3. Polyuria
4. Polydipsia
5. Sudden weight loss
6. Weakness
7. Polyphagia
8. Genital thrush
9. Visual blurring

· Itching

· Irritability

· Delayed healing

· Partial paresis

· Muscle stiffness

· Alopecia

· Obesity

Output

1. Predicted Diabetes Status
2. Negative (0) – No diabetes
3. Positive (1) – Diabetes detected

Tags:

#ADSTLSTM #DiabetesPrediction #DiabetesRiskAssessment #DeepLearning #LSTMNetwork #TemporalModeling #HealthcareAI #MedicalDataAnalytics #ClinicalDecisionSupport #AIinHealthcare #PredictiveModeling #MachineLearning #DigitalHealth #IntelligentSystems #ExplainableAI #HealthInformatics #MedicalAI #ChronicDiseasePrediction #TechInnovation

Reference

 Chen, Y., et al. “Quantum Dilated Convolutional Neural Networks.” IEEE Access (2021).

 Matondo-Mvula, N. & Elleithy, K. “Breast Cancer Detection with Quanvolutional Neural Networks.” Entropy (2024).

 “Modified Quantum Dilated Convolutional Neural Network for Cancer Prediction Using Gene Expression Data.” PubMed