The Research And Application On Physical Examination Data Preprocessing Methods

The medical physical examination data has accumulated rich and valuable information, it can be used to analyze the risk of disease and personalized health guidance and to predict the risk and probability of some chronic diseases, and to remind the subject to discover potential disease in time, to provide health guidance and disease treatment measures for Physical examination person. But, the original physical examination data has many problems, including ambiguity, noise,incomplete and redundancy information, so it cannot be used for disease risk assessment and prediction directly. Therefore, it is very important to handle the data of medical examination.In order to solve this problem and make full use of the valuable information in the data, several methods are proposed in this thesis. A compression-based data reduction method is used to reduce the time and space complexity of the data in order to solve the problem of information redundancy; a data cleaning based on the similar duplicate records and missing value is used to complete the data cleaning and solve the problem of inconsistency and solve the problem of the non-standard data and outliers, duplicate values and missing value of physical examination data, By removing a tuple, ignoring the incomplete data and filling technology to complete value data cleaning of the missing; a data transformation method based on linear function is used to get the consistency and continuity of the history data and solve the missing of unique identification code. Finally, this thesis puts forward the innovational work that a field matching algorithm based on segmentation and weights completes the detection of similar duplicate records. The purpose of data preprocessing in physical examination is that the non-standard data can be converted into standard data, it can unified the doctor terms and medical conclusion and correct mistakes and fill the vacancy information.The experimental results show that a compression-based data reduction method can reduce the irrelevant and redundant information of physical examination datagreatly. The field matching algorithm based on segmentation and weights was higher than that of traditional algorithm 6.23% and 5.44% and 5.84% in the recall ratio,accuracy and F-measure values, it improved that the accuracy of algorithm is higher than the traditional algorithm in detecting similar duplicate records. A data transformation method based on linear function can add unique identification code successfully.Finally, a physical examination data query system is developed to realize the query of physical examination data, and visual display of the physical examination data.