Research On Visual Analysis Of Electronic Medical Obstetric Examination Data

The modern electronic medical record system uses a large number of text charts to display the physical examination data of pregnant women,but it wastes time,which leads to doctors unable to find the required information quickly and to diagnose the physical condition of pregnant women in time.At the same time,although the indicators of the pregnant women during the labor examination indicate that their physical condition is good,but only single indicators can not judge the potential diseases of the pregnant women themselves,such as pregnancy complications caused by other reasons during the pregnancy or not,how to describe the health of pregnant women with multiple indicators of the labor examination is an important research topic.In this paper,the electronic medical records of pregnant women during pregnancy examination in hospitals were used,that is,the basic information of pregnant women during pregnancy examination and the records of routine gynecological examination items.After preprocessing the pregnancy test data,the paper used a variety of clustering algorithms to separate the labor test data between clusters,aiming to help doctors find the role of the key labor test factors of pregnant women in each cluster on pregnancy complications.At the same time,the case sample similarity algorithm was proposed to seek similar cases of pregnant women seeking treatment in the same cluster and analyze their potential pregnancy complications.Finally,this paper proposes a medical visualization system based on clustering.The main contribution of this paper is reflected in the following three aspects:1.To get to the actual medical significance of cluster,explore different key prenatal indicators in each cluster,mining can identify the pregnancy complications but neglected prenatal index,based on the optimization based on distance measure of rob minkoff algorithm,is put forward based on multidimensional the indicators of a similarity measure algorithm,cluster generate more reasonable of pregnant women.The comparison of visual analysis effect map and the value of quality evaluation index prove that it can divide pregnant women cluster more effectively.Compared with the traditional similarity measurement algorithm based on distance measurement,the multi-dimensional similarity measurement algorithm is more referential and practical.2.In order to identify the correlation of the factors of key examination items in pregnant women,similar case samples of visiting pregnant women were searched to better analyze the potential pregnancy complications similar to those in the case samples during pregnancy.In this paper,the case sample similarity measurement algorithm is used to find the approximate case and narrow the range of comparison.By comparison with experimental data random sampling,is said to this algorithm the most similar cases sample type of pregnancy complications compared with the individual’s actual pregnancy complications type whether its consistent proportion gain value size,a size ratio value score is higher,so that the algorithm can provide clinic medical diagnosis of pregnant women with more detailed case reference.3.A visual analysis system for identifying potential pregnancy complications was designed and implemented in this paper based on the large-scale data of routine gynecological and obstetrical examination of pregnant women,the similarity measurement algorithm of multidimensional obstetrical examination indicators and the visual system based on cluster analysis.The system integrates a number of medical information data visualization technologies and visual analysis functions to analyze potential pregnancy complications of visiting pregnant women.Finally,three kinds of experimental cases are used to prove the effectiveness and practicability of the visualization system.