A Prediction Model for the Risk of Osteoporosis Fracture in the Elderly Based on a Neural Network

摘要

A prediction model for the risk of osteoporosis fractures in the aged people is proposed based on a neural network system. A back propagation (BP) neural network is used as the key component in the prediction model. In our prediction model, not only the bone mineral density (BMD), but also some conventional medical examination data from renal function test (RFT), routine blood test (RBT) and liver function test (LFT) are considered. About 5,000 sample data are extracted from more than 20,000 cases in the Seventh People's Hospital of Chongqing, China. The BP neural network is trained by a 10-fold cross validation method. Then, the well trained BP neural network is used to predict the risk of osteoporosis fractures. Moreover, we also test the performance of the proposed model in the condition of privacy preservation by hiding gender and age. The experiment results show that the prediction accuracy of the proposed model is more than 70%. Therefore, the model has the high potential to auxiliary diagnose the possibility of osteoporosis fractures.