假设检验

t 检验、F 检验、卡方拟合优度检验等

Statistics and Machine Learning Toolbox™ 提供参数化假设检验和非参数化假设检验，帮助您确定样本数据是否来自具有特定特征的总体。

分布检验（如安德森-达林检验和单样本科尔莫戈洛夫-斯米尔诺夫检验）可以检验样本数据是否来自具有特定分布的总体。双样本科尔莫戈洛夫-斯米尔诺夫检验可以检验两组样本数据是否具有相同的分布。

位置检验（如 z 检验和单样本 t 检验）可以检验样本数据是否来自具有特定均值或中位数的总体。双样本 t 检验或多重比较检验可以检验两组或多组样本数据是否具有相同的位置值。

散度检验（如卡方方差检验）可以检验样本数据是否来自具有特定方差的总体。双样本 F 检验或多样本检验可以比较两个或多个样本数据集的方差。

通过交叉表分析和随机性游程检验确定样本数据的其他特征，并确定假设检验的样本大小和幂。

函数

全部展开

分布检验

`adtest`	Anderson-Darling test
`chi2gof`	Chi-square goodness-of-fit test
`crosstab`	Cross-tabulation
`dwtest`	Durbin-Watson test with residual inputs
`fishertest`	Fisher’s exact test
`jbtest`	Jarque-Bera test
`kstest`	One-sample Kolmogorov-Smirnov test
`kstest2`	Two-sample Kolmogorov-Smirnov test
`lillietest`	Lilliefors test
`mmdtest`	Two-sample multivariate hypothesis test using maximum mean discrepancy (MMD) (自 R2024b 起)
`runstest`	Run test for randomness

位置检验

`friedman`	Friedman’s test
`kruskalwallis`	Kruskal-Wallis test
`multcompare`	Multiple comparison test
`ranksum`	威尔科克森秩和检验
`sampsizepwr`	Sample size and power of test
`signrank`	Wilcoxon signed rank test
`signtest`	Sign test
`ttest`	单样本和配对样本 t 检验
`ttest2`	双样本 t 检验
`ztest`	z 检验

散度检验

`ansaribradley`	Ansari-Bradley test
`barttest`	Bartlett’s test
`sampsizepwr`	Sample size and power of test
`vartest`	Chi-square variance test
`vartest2`	Two-sample F-test for equal variances
`vartestn`	Multiple-sample tests for equal variances

估计统计量

`meanEffectSize`	One-sample or two-sample effect size computations (自 R2022a 起)
`gardnerAltmanPlot`	Gardner-Altman plot for two-sample effect size (自 R2022a 起)

批量漂移检测

检测漂移

detectdrift Detect drifts between baseline and target data using permutation testing (自 R2022a 起)

访问测试结果

DriftDiagnostics Diagnostics information for batch drift detection (自 R2022a 起)

检查测试结果

`summary`	Summary table for `DriftDiagnostics` object (自 R2022a 起)
`ecdf`	Compute empirical cumulative distribution function (ecdf) for baseline and target data specified for data drift detection (自 R2022a 起)
`histcounts`	Compute histogram bin counts for specified variables in baseline and target data for drift detection (自 R2022a 起)
`plotDriftStatus`	Plot p-values and confidence intervals for variables tested for data drift (自 R2022a 起)
`plotEmpiricalCDF`	Plot empirical cumulative distribution function (ecdf) of a variable specified for data drift detection (自 R2022a 起)
`plotHistogram`	Plot histogram of a variable specified for data drift detection (自 R2022a 起)
`plotPermutationResults`	Plot histogram of permutation results for a variable specified for data drift detection (自 R2022a 起)

主题

Available Hypothesis Tests
View hypothesis tests of distributions and statistics.
Hypothesis Testing
Hypothesis testing is a common method of drawing inferences about a population based on statistical evidence from a sample.
Hypothesis Test Terminology
All hypothesis tests share the same basic terminology and structure.
Hypothesis Test Assumptions
Different hypothesis tests make different assumptions about the distribution of the random variable being sampled in the data.

精选示例

Selecting a Sample Size

Determine the number of samples or observations needed to carry out a statistical test. It illustrates sample size calculations for a simple problem, then shows how to use the sampsizepwr function to compute power and sample size for two more realistic problems. Finally, it illustrates the use of Statistics and Machine Learning Toolbox™ functions to compute the required sample size for a test that the sampsizepwr function does not support.

打开脚本