Saudi Journal of Oral Sciences (Dec 2024)

A decision support system based on classification algorithms for the diagnosis of periodontal disease

  • Abdulrahman Alshehri,
  • Mohammed Dahman,
  • Mousa Assiri,
  • Abdulkarim Alshehri,
  • Sharifah Alqahtani,
  • Mohammed Shaiban,
  • Bashyer Alqahtani,
  • Sabah Althbyani,
  • Hatem Alhefdi,
  • Khalid Hakami,
  • Abdulbari Ali,
  • Abdullah Saeed

DOI
https://doi.org/10.4103/sjoralsci.sjoralsci_50_24
Journal volume & issue
Vol. 11, no. 3
pp. 181 – 192

Abstract

Read online

Introduction: Inflammatory conditions, including gingivitis and periodontitis, affect the supporting structures of teeth. The early detection of these diseases is critical for the prevention of systemic complications. Recent advances in artificial intelligence have introduced novel diagnostic methods that offer the potential for more accurate and personalized diagnostics. Aims: The purpose of this study was to develop and evaluate a decision support system (DSS) based on selected classification algorithms, namely random forest (RF), support vector machine (SVM), artificial neural network (ANN), and logistic regression for the periodontal disease diagnosis. Materials and Methods: A cross-sectional study design was adopted, A 300 patients were taken, split into a training set with 80% and a test set with 20%. The algorithms were used to analyze 19 demographic, clinical, and radiographic parameters. Chi-square and ANOVA tests were conducted and for these, a significance level of 0.05 was used. We calculated accuracy, precision, recall and F1-score, which we used to assess model performance. Results: The RF and ANN models exhibited very good performance, reaching almost perfect accuracy, precision, recall and F1 scores, showing strong potential as a diagnostic. Nevertheless, the SVM and decision tree models equally yielded robust results with a balanced generalizability across metrics. Results showed little effectiveness for the Naïve Bayes on grading periodontal disease. Discussion and Conclusion: DSS models show promise in enhancing periodontal disease diagnostics, particularly RF and ANN models. However, the risk of overfitting and limited performance of simpler models, such as naive Bayes, underscores the need for further research, including external validation, to ensure the reliability and generalizability of these models in clinical practice.

Keywords