Digital Decision Making In Dentistry: Analysis And Prediction of Periodontitis Using Machine Learning Approach

Abstract

Machine Learning and Deep Learning, the powerful weapons of Artificial Intelligence plays crucial role and major contributions in almost all the sectors using sophisticated technological framework with an insight to unlock the needs of business enacting decisions where data patterns holds a main role right from data gathering and explorations to visualization and predictions. Recently it is also a predominating technology used vigorously in various health sectors like medical, dental and allied health services as an aid to develop tools for decision making in data analytics and exploration, disease prediction and control, data analytics and treatment planning. Major research on this area was done in past contributing good frameworks for the predictions of Breast cancer, Rheumatoid Arthritis, Osteoporosis, Diabetes, Sarcoidosis, Graves’ disease, AIDS, Psoriasis and many more. The current research paper is a result of using such Machine Learning approaches for the prediction of Periodontitis, a most common gum disease which leads to severe complications like tooth supporting structure loss like bone loss around tooth, ligament loss and finally the tooth loss if left untreated. In the current paper, a dataset of 206 sizes of diabetic and non-diabetic periodontitis patients were collected with the measurable parameters like age, sex, oral hygienic status, tooth mobility, periodontal index, gingival index, furcation, alveolar bone loss, pocket depth and other parameters were taken and implemented. Supervised machine learning algorithms for classification like Support vector machine, Naïve Bayes, Random forest, Logistic regression, decision tree and K Nearest neighbor algorithms were used and implemented in python using jupyter notebook for the prediction of periodontitis and obtained accuracies of 96.7%, 95.1%, 96.7%, 93.5%, 96.7% and 98.3% respectively. The current paper demonstrates how the dataset was collected and implemented using Machine learning approach in dentistry for obtaining a suggestible predictable model for periodontitis also compares all the models mentioned for their efficiency and accuracies.