# Sample data ,35 The blood glucose concentration of a healthy man before eatingĭf = pd.DataFrame(data, columns =)Į = df.mean() # Calculate the mean
Judge whether it conforms to normal distribution PS：t- The hypothesis of the test is that the test data satisfy the normal distribution, Otherwise, for the small sample data which does not satisfy the normal distribution, use t- The test will cause a big deviation, Although for the large sample data which does not satisfy the normal distribution t- Testing is also a fairly accurate and effective means. It is very common to analyze whether there are differences between two groups of data. When the sample size is small ,KS The test is the most nonparametric test. Of course, the cost of such convenience is when the tested data distribution conforms to a specific distribution ,KS The sensitivity of the test is not as high as that of the corresponding test. KS Test and t- The difference between other methods such as testing is KS The test does not need to know the distribution of the data, It can be regarded as a nonparametric test method. KS Test is a statistical test method, By comparing the frequency distribution of the two samples 、 Or the frequency distribution of a sample is different from a specific theoretical distribution （ Like a normal distribution ） To infer whether the two distributions come from the same distribution. Ttest,pval = ttest_ind(s1,s2,equal_var=equal_var) '''t Test whether there are differences between the two population mean values represented by independent samples ''' '''F Test whether the variance of the sample population is equal ''' In addition to requiring samples from normal distribution, The population variance of the two samples is also required to be equal “ Homogeneity of variance ”. Print("Null Hypothesis:mean(s1)=mean(s2),α=0.05")įor the third question, independent samples t test, Compare whether there is a significant difference between the two population mean values represented by the two samples. The difference of paired samples can be used as a variable, The overall mean of the difference is 0, It follows a normal distribution. Paired samples are mainly the comparison of effects before and after the same experiment, Or the comparison of the test results of the same sample by two methods. from scipy.stats import ttest_1sampĪges = The population obeys normal distribution, Take a sample from a normal population n A sample of individuals, Calculate the mean and standard deviation of the sample, Determine whether the mean value of the population is the same as that of the sample. Single sample t The test is the comparison between the sample mean and the population mean. We refer to 《python Scientific computing Second Edition 》： T Test is a statistical analysis method suitable for small samples, By comparing the mean values of different data, Study whether there are differences between the two groups of data. In a model, An important technique is to determine whether the characteristics of the training set and the test set are identically distributed, This is also a very important assumption of machine learning, But many times we acquiesce to this truth, But it is difficult to have a way to ensure that the data are uniformly distributed.
0 kommentar(er)
