Bootstrap And Its Application In Laboratory Proficiency Testing

Laboratory proficiency testing is an activity to evaluate the detection capability of labs, by analyzing and contrasting the experimental data of labs. With the progress of society, proficiency testing provides an objective measure for the participants evaluating the reliability of the experimental results and it is a valid certification of participant labs’ testing capability. There are two main bodies, the organizers and the participants, in each proficiency testing. The most authoritative organizers in the domestic called CNAS, and units that are qualified, such as inspection and quarantine bureau in provinces and cities, also can play the role of organizer. The participants are usually those laboratories that being accepted and have been accepted. The processes of proficiency testing are in the below order:the organizers organize and design the proficiency testing, prepare samples and distribute to participants, recycle data, determine the specified value and the standard deviation, determine the performance statistics to assess lab ability, report the results to the participants, handling opinions.At present, traditional robust statistical method is a common statistical method in proficiency testing. Its premise condition is that the original data should obey the normal distribution. However, the data from participants in the practical work with the characteristics as less sample size, unknown distribution and so on. In this case, using traditional methods will lead an accurate result and impact the quality of the proficiency testing. The traditional Z value is up to specified value and the standard deviation which are determined by all participants’ lab data. Therefore, the length of time becomes an important factor that affect the work efficiency of proficiency testing.As is known to all, incremental analysis and bootstrap is often used to deal with data that is normal distribution and bootstrap apply to small samples. Therefore, the key research of this paper is that how bootstrap can be used in proficiency testing. At the same time, in order to provide a service platform for the participants and the society, improving the efficiency of proficiency testing is inevitable. By simulating a large number of examples of data and analysis, the paper give a program to improve the efficiency by using the principle of sequential and bootstrap.This paper is divided into four chapters:chapter one is a introduction about the background and signification of proficiency testing, the status of proficiency testing at home and abroad and the main contents; the second chapter briefly introduce the related concepts and problems of proficiency testing, traditional robust statistical methods and the principle of bootstrap; the third chapter focuses on the application of bootstrap in proficiency testing by theoretical analysis, simulation analysis and empirical analysis;the fourth chapter analyzes and summarizes research findings, discuss future research directions.