Artificial Intelligence-Based effective healthcare prediction system for diabetic patients

Diabetes, a chronic disorder characterized by elevated blood sugar levels, requires a high degree of care for diabetic patients. Diabetes treatment requires frequent monitoring, medication management, nutritional control and lifestyle changes to prevent complications. Effective diabetes treatment assists diabetic individual's healthier lives by reducing the risk of heart disease, kidney difficulties and visual problems. Healthcare providers are essential in instructing and helping patients as they struggle to manage this disease and preserve their general health. In this study, we proposed the Tabu Search-optimized Bidirectional Long Short-Term Memory (TS-BiLSTM) model to enhance the predictive accuracy of healthcare systems for diabetic patients. We gathered datasets from the National Institute of Diabetes Diseases from Kaggle open source, which includes glucose, age, BMI, etc. Initially, we applied min-max normalization to preprocess the collected data and utilized Principal Component Analysis (PCA) to extract key features from the preprocessed data. To improve the efficiency of the BiLSTM, we implemented the TS optimization technique. This combination of preprocessing and optimization strategies improves the BiLSTM model's performance and accuracy in evaluating and analyzing data. The result analysis phase is done on various parameters such as accuracy (95.72%), precision (94.5%), recall (89.68%) and F1-score(90.02%). The findings are compared with other existing methods and our suggested TS-BiLSTM approach performed better than different current methodologies.