Classification and Prediction of Oral Diseases in Dentistry using an Insight from Panoramic Radiographs and Questionnaire

The healthcare domain is a very important research field with rapid technological advancement, where programmed medical assistance systems or mathematical devices are high in demand to utilize the advances in research in Machine Learning (ML). Comparing with rapidly growing artificial intelligence research in other healthcare domain and dental care is relatively slow in dental care. This paper will discuss predictive analytics in the specialized field of oral health care, i.e. Dentistry, which is considered as a branch of medicine dealing with anatomy, development, and diseases of teeth. The emphasis is given to this field because it is a difficult mission, due to short of precise data availability in dentistry for diagnosis and prediction. Predictive analytics help practitioners to take sensible decisions connecting to the patient's physical condition and cure. In this paper, an approach towards the diagnosis, prediction, and classification of oral health diseases in dentistry is implemented and discussed with the help of the machine learning techniques on the self-created dataset of 133 patients. Now a day's panoramic dental x-ray images are an essential diagnostic tool to detect the symptoms at an early stage by dentist, but always exposure to radiation for every patient is not possible so at some extend we have to avoid it. So to solve this issue diagnosis at early stage is very important. Early diagnosis can be done on the basis of symptoms of patient. To detect the possibility of disease (in this study tooth wear) is considered at an early stage, for which one questionnaire is created which is collection of data from different patients on the basis of symptoms. Mainly tooth wear is the damage of tooth tissue and structures not due to caries. It can happen in numerous forms either attrition, abrasion, erosion, non-caries cervical lesion (NCCL) or an amalgamation of two or more forms. This questionnaire contains all the possible parameters for occurring diseases are considered. Here classification and prediction is done for four different classes (class 1, 2,3and 4) based on severity ratio (class 1-Low severity and class 4-High severity). To perform the classification and prediction on the questionnaire which is in CSV format support vector machine is used. The objective of this study is to determine the patterns and associated etiologies of tooth wear among adults for Indian population.