《学术论文集锦[开错]》紫微客

10、010

　　4.3 Defects and Improvements of innovative method

　　As we known, the classical statistical methods used in the experimental process in Part 4.2 is based on the law of large numbers, which determines that when reliability analysis is carried out, sufficient sample size is needed to estimate the characteristics of unknown distributions. However, if the sample size is small, when judging the probability of uncertain events, most people tend to transplant the law of large numbers to small samples, for example, mistakenly thinking that the mean of small samples also tends to the mean of the population. This behavioral deviation is called the law of small numbers by Kahneman and Tversky.

　　Therefore, we can make some improvements for the above experimental process because the data set in this project is small sample, that is, to adopt statistical analysis methods suitable for small samples, such as the innovative design based on Bootstrap method and Bayes theory.