1,2Noorul Islam Centre for Higher Education, Tamilnadu, India
Contact details: annesheaba@gmail.com
Background: Due to their impaired bone metabolism and glycemic instability, diabetic patients who undergo dental implant procedures run a very high risk. This research work presents a novel framework for Dental Implant recommendation for diabetic patients. For diabetic patients, this framework suggests the best implant types and individualized preoperative strategies. This framework integrates a synthetic patient data modelling with more clinically significant features like HbA1c, bone density and glycemic control indicators. 3000 data which mimics the clinical data is generated and with which the model is trained.
Methods: The Particle Swarm Optimization (PSO), which imitates the behavior of swarm of birds, is used to optimize the features. For the purpose of finding the best set of solutions, elitism is used. The classification is done by using Support Vector Machine. With a maximum classification of 81% and an F1-weighted score of up to 0.92, this combination demonstrated strong performance. The implant types, such as Zirconia, were also subjected to the ROC analysis, which yielded AUC scores greater than 0.90, confirming the proposed framework's discriminatory power. Additionally, the dynamic generation of the clinical recommendations for implant timing and glycemic management was used.
Conclusion: Based on these findings, the proposed framework can function as a patient-specific, intelligent, and comprehensible decision support tool for dental implant planning in diabetic care.
Keywords: Nature Inspired Computing, Dental Implant Recommendation, Predictive Science, Disease Prediction.
How to cite this article: Sheaba AGS, Anitha A. Data-Driven Dental Implant Recommendation System for Diabetic Patients Using Particle Swarm Optimization and SVM. Int J Drug Deliv Technol. 2026;16(12s): 476-485. DOI: 10.25258/ijddt.16.12s.59
Source of support: Nil.
Conflict of interest: None