The Applications Study Of Data Mining Technology In The Data Analysis Of The Neonatal

The raw data collected by hospital are increasing year by year, so a large number of raw data have been stored, including patients'basic information and various cases. Huge useful knowledge needs to be mined under the increasing data. How to extract and mine the knowledge are currently hot topics. As an effective method of information extracting and knowledge discovery, data mining extracts useful information from medical databases and discovers the potential hidden data model by analyzing and evaluating these data. It provides scientific judgments and medical treatment, which helps people improve the understanding of fertility and enhances the research and management of modern fertility.In recent years, the Shanghai Health Bureau took the leading in establishing a database of newborn's birth status. It would help a lot in reasonable distribution of health facilities and resources and prevention of neonatal diseases by counting and mining these neonatal data. This paper took Shanghai neonatal data as an example and did the research and application of data mining technology on neonatal data analysis. It provided the quantitative basis for prenatal and postnatal care through the analysis various factors that impact the birth of newborn. It facilitated the decision makers to understand the neonatal status in each district of Shanghai. It helped them to make right decisions and provided a reference for expert diagnosis.Firstly, this paper introduced the data pretreatment of newborn, such as data cleaning, data conversion and dimension reduction. Through these methods, it effectively processed the missing data, noise data and inconsistent data. Data pretreatment improved the quality of data and the knowledge amount gained by data mining.Then, this paper introduced the algorithm of mining association rules, focusing on the classical algorithm Aprior algorithm. It proposed the advantages and disadvantages of Aprior algorithm. Focusing on the disadvantage that Aprior algorithm needed to repeatedly scan the database and produced a large number of frequent item sets, this paper proposed an improved Aprior algorithm based on division to analyze the association rules of neonatal data.Finally, the paper designed the Web-based Visual Data Mining System of Newborn, which considered more about the interactions between users and system. It used rich and interactive charts provided by Flex technology to display a very intuitive and effective data analysis results and made description and implementation of each steps. By analyzing raw data of a large number of newborns stored in Shanghai Hospital, it got some potential birth rules, which facilitated the decision-makers'real-time monitoring of neonatal status and trends in each district of Shanghai, which had a certain practical significance.