Abstract
Jaw fracture is one of the common facial injuries in human body. The jaw fracture disease is one of the tenth most injuries in human body. Jaw fracture may result from high physical injuries impacts and medical conditions that weaken the looseness of the teeth, jaw stiffness, swelling and so on. Due to the various symptoms of this disease, diagnosis is difficult at an early stage. Therefore, different machine learning and data mining techniques are used to detect jaw fracture because of the severe effects of the disease. Initially, jaw related data is randomly collected from the patient, which includes information such as the patient name, age, height, weight, osteoporosis risk factors, prediction, fracture level, and so on. After collecting jaw fracture information, unwanted data is discarded by applying normalization techniques; optimized features are then selected using the metaheuristic firefly algorithm. The selected data is then classified using the multi-layer neural network training based on associative neural networks. The classifier then successfully categorizes the abnormal jaw fracture feature. Then, the efficiency of the system is examined with the help of the mean square error rate, precision, recall, and accuracy.