The Study Of Some Problems In Medical Imaging Data Warehouse System

In recent years, the medical imaging department has accumulated a large number of businessdata. On the one hand, how to effectively search the information for research and teaching hasbecome the focus of clinicians’ concern. So research on automatic classification and full-textretrieval technology to improve the traditional database sutrctured query is very important. On theother hand, how to build the data warehouse, analyze multidimensional data, and develop the datamining application, as well as auxiliary decision-making have become the focus of the leader’sconcern. On the above requirements, this paper began some researches as follows:1. Medical equipment performance research based on GM(1,1) grey forecasting model: Atfirst, build the data warehouse, adopt the medical equipment performance key indicators methodand radar chart method to realize the equipment exam quantity OLAP analysis and provide theintuitive data display; Second, set up equipment exam quantity model based on GM(1,1) to realizeexam quantity prediction. Third, establish the profit and loss balance analysis model to realizeequipment efficiency analysis and provide the decision makers intuitive and deep statistical chartdata.2. Research on disease automatic classification based on naive bayes algorithm: We study thetwo issues including medical diseases automatic classification and medical examination reportclassification retrieval in-depth. Use the k-means algorithm to cluster the disease information ofmedical examination report, then use the improved naive bayesian classifier to realize the accurateclassification of the disease, finally use.Net technology to realize the medical examination reportinformation classification retrieval.3. Research and application on the full-text retrieval technology based on inverted index: Atfirst, discuss the information retrieval, Lucene search engine, Chinese word segmentationtechnology and its application performance testing, and two modes of positive index and invertedindex. Second, on the analysis of the difference between the positive index and retrievalperformance, this chapter mainly study the query optimization techniques based on the invertedindex structure. Third, use the.net technology design and implement an application system whichuse full-text retrieval of the medical disease information and the examination report, from thereport related to medical consultation table and the follow-up table, so as to assist doctors inmedical diagnosis. At last, embed the full-text retrieval application in medical image datawarehouse system, contrast and analysis of the pros and cons between the combination query andthe full-text retrieval tecnolgies, therefore effectively improve the efficiency of the doctor’sscientific research.