Research On Cardiovascular Disease Prediction Based On Deep Learning Technical

With the development of computer technology, the computer technology applied in medical treatment for diseases forecasting has become a focus of research, and has very important significance. Deep learning technology is a latest research results in machine learning area in recent years, this thesis attempts to apply deep learning technology to the disease forecasting of cardiovascular disease.The main works of this thesis is as follows:(1)Data normalization of cardiovascular diseasePointing at the dimensional effect caused by the physical significance difference and orders magnitudes difference of the human body physiological data’s attribute values, this thesis used the Z- score standardization method to calculate the mean value of each data set’s properties and data set’s standard deviation, accomplishing the data set’s normalization.(2)Based on DBN, modeling the cardiovascular disease forecastingCombined with the data characteristics of cardiovascular disease, build the prediction model of cardiovascular disease based on DBN. The training core of the model include restricted Boltzmann machine network and BP neural network, using the greedy layer-layer learning algorithm to train and then to confirm the input node, layers of hidden layer and the node number of hidden layer, finally work out the best DBN prediction model structure. Experiment shows that this model has a better accuracy compared with BP neural network.(3)Design and implement the cardiovascular disease prediction system prototypeUsing the constructed DBN model, cardiovascular disease prediction system prototype was designed and implemented. The system mainly includes three modules: data preprocessing module, cardiovascular disease prediction module, prediction display module, which provide a prediction method for cardiovascular diseases.