Class TTest

java.lang.Object
cloud.opencode.base.math.stats.inference.TTest
public final class TTest extends Object
Student's t-test implementations for hypothesis testing. Student t 检验实现，用于假设检验

Provides one-sample, independent two-sample (Welch's), and paired t-tests. All methods return a TestResult with two-tailed p-values.

提供单样本、独立双样本（Welch）和配对 t 检验。 所有方法返回包含双尾 p 值的 TestResult。

All methods are stateless and thread-safe.

所有方法无状态且线程安全。

Since:
JDK 25, opencode-base-math V1.0.3
Author:
Leon Soo www.LeonSoo.com
See Also:

  • Method Details

    • oneSample

      public static TestResult oneSample(double[] data, double mu0)
      Performs a one-sample t-test (H0: population mean = mu0). 执行单样本 t 检验（H0: 总体均值 = mu0）

      Computes t = (mean - mu0) / (s / sqrt(n)), df = n - 1, two-tailed p-value.

      计算 t = (mean - mu0) / (s / sqrt(n))，df = n - 1，双尾 p 值。

      Parameters:
      data - the sample data (at least 2 elements) / 样本数据（至少 2 个元素）
      mu0 - the hypothesized population mean / 假设的总体均值
      Returns:
      the test result / 检验结果
      Throws:
      MathException - if data is null, has fewer than 2 elements, or contains NaN/Infinity

    • twoSample

      public static TestResult twoSample(double[] x, double[] y)
      Performs an independent two-sample t-test using Welch's method (unequal variances). 使用 Welch 方法执行独立双样本 t 检验（不等方差）

      Computes t = (meanX - meanY) / sqrt(sX^2/nX + sY^2/nY), with Welch-Satterthwaite degrees of freedom approximation.

      计算 t = (meanX - meanY) / sqrt(sX^2/nX + sY^2/nY)， 使用 Welch-Satterthwaite 自由度近似。

      Parameters:
      x - first sample data (at least 2 elements) / 第一组样本数据（至少 2 个元素）
      y - second sample data (at least 2 elements) / 第二组样本数据（至少 2 个元素）
      Returns:
      the test result / 检验结果
      Throws:
      MathException - if data is null, has fewer than 2 elements, or contains NaN/Infinity

    • paired

      public static TestResult paired(double[] x, double[] y)
      Performs a paired t-test (H0: mean difference = 0). 执行配对 t 检验（H0: 平均差值 = 0）

      Computes differences d = x - y, then performs a one-sample t-test on d with mu0 = 0.

      计算差值 d = x - y，然后对 d 执行 mu0 = 0 的单样本 t 检验。

      Parameters:
      x - first sample data / 第一组样本数据
      y - second sample data (same length as x) / 第二组样本数据（与 x 等长）
      Returns:
      the test result / 检验结果
      Throws:
      MathException - if arrays are null, different lengths, have fewer than 2 elements, or contain NaN/Infinity