Research And Application Of Visualization Algorithms Based On Medical Big Data

With the rapid development of "Internet+",medical big data has many data types,complex relationships and explosive growth.It is ineffectively displayed by the com-mon visualization methods.The medical big data visualization technology has many challenges.The existing visual methods can't meet a large variety of medical big data and complex high-dimensional needs;visual analysis results unsatisfactory.The exist of medical big data point out research directions.It has contributed to medical.The main contents of this paper are as follows:1.The proposed medical visualization of large data classification methods:This paper outlines the origin,characteristics and research progress of medical big data,introduces the related concepts and research status of visualization of medical big data.The existing visualization methods of medical big data are studied,which give a comprehensive description of the characteristics,legend and common visualization methods of medical big data by using a table.2.The propose of a composite dynamic multi-class decision-making radar(DMDR)method:The method of dynamic classification decision radar is proposed to extend the traditional Time Radar Tree method.By taking a multiple attribute single graph as a single graph class,many single graphs in a class axis are integrated into N-dimensional dynamic planar graph.Finally,the weighted dependencies of multiple attributes in a single graph is dynamically visualized.This method not only reduces the visual clutter caused by multidimensional irregular data,but also dynamically displays the hidden information between the key features of the complex data.3.Based on GBDT + LR fusion model to improve the assessment of visual analysis of the results:Medical big data to Visualization of Individual Physical Sub-Health by GBDT&LR Fusion Models.It is better than Logistic Regression model.The experiment have more accurate prediction accuracy and stability.GBDT&LR fusion model proposed to evaluate the medical big data with complex structure.Such as high-dimension and multi-attribute.Abnormal data and key features of real-time dynamic display,so provide both doctors and patients more reliable auxiliary treatment.