Independent two-samples test in Matlab

Yang Runhuai

1

楊潤懷(1987-)，男，講師，生物3D打印

Zhang Zhen

1

Yang Siqiao

1

Liang Zhen

1

梁振(1981-)，男，副教授，臨床工程

1、Life Science School, Anhui medical university, hefei, 230032

Abstract：Purpose: Based on the principle of independent two-samples test and code it in matlab. The code will automatically select the statistical method and calculate statistical results. Method： Firstly，the program judges the two groups of samples whether obey the normal distribution; If any sample does not obey the normal distribution,the K-S and M-W U test were used to test whether distribution of the samples are the same; If both of the two samples obey the normal distribution, then uses the F test to judge whether the two samples have equal variance. if the variance is equal, then uses the t test directly; If the variance is not equal, then uses Satterthwaite to calibration and uses t test. Results: we achieved the above functions in matlab, and the two instance confirm the rightness of the code. Conclusion: This paper provides a method which can automatically chooses the statistical method for the independent two-samples test, and give reliable statistical results.