| With the development of medical data analysis techniques, the strong demand from clinical treatment, research learning makes it become a hot topic. How to find an effective prevention method on occupational diseases becomes an urgent problem to solve. Classical data analysis methods are widely used for diseases with low forecast accuracy. At the same time, hospitals have a large number of cases of clinical data every day, which makes the sharing and using of the whole information extremely low and difficult.According to these circumstances, the medical subjective analysis system (MedicalSAS) provides a professional medical analysis platform for medical experts, with the combination of medical analysis and the sharing of resources, job scheduling on grid. Modelling on the submitted disease data through effective data analysis algorithm, it can predict the typical characteristics of disease, and give suggestions on the analysis of disease information. One of the key technologies in MedicalSAS is the establishment of combined forecasting methods on the silicosis and hypertension diseases. Two combined forecasting methods based on BP neural network are proposed to predict the possible rate and ages of sufferers to suffer the diseases. At the same time, CGSP 2.0 is used to implement the encapsulation and deployment of data analysis grid services, and the operations on data preparation, model establishment, model evaluation and model visualization are also provided. The platform is user-friendly and scalable.Function and performance test showed that, the combined forecasting methods based on the traditional data analysis algorithms can make a prediction to exert the superiorities of the time series datum of dust-exposed workers and other pathogenic factors, and the efficiency and accuracy of the hybrid models are enhanced greatly contrasted with single BP neural network. The hybrid models can be effective methods for the silicosis and hypertension diseases prediction. |
PDF Full Text Request