statkit API

statkit 是纯 Dart 统计内核包，路径 packages/statkit，可独立发布到 pub.dev。

安装

dependencies:
  statkit:
    path: packages/statkit
    # 或: ^0.1.0

快速示例

import 'package:statkit/statkit.dart';

void main() {
  final data = [2.1, 2.5, 2.8, 2.3, 2.6];

  final d = Descriptives.compute(data);
  print('mean=${d.mean} sd=${d.sd}');

  final t = TTest.oneSample(data, mu0: 2.0);
  print('t=${t.t} p=${formatP(t.pTwoTail)}');
}

API 一览

Descriptives

Descriptives.compute(List<double>) → Descriptives

字段：n mean sd variance min max range median q1 q3 iqr skewness kurtosis se sum sumSquares ciLower95 ciUpper95 sorted

辅助：percentile(sorted, p)、modes(values)

TTest

• oneSample(data, {mu0})
• independentSamples(g1, g2, {autoWelch})
• paired(a, b)

OnewayAnova

compute(groups, {labels}) → AnovaResult
含 f p ssBetween ssWithin etaSquared omegaSquared levene groups

ChiSquareTest

• goodnessOfFit(observed, {expected})
• independence(observedMatrix)

Correlation

• pearson(x, y) / spearman(x, y)
• matrix(columns) / nMatrix(columns)

Regression

• simple(x, y, {xName, yName})
• multiple(xs, y, {predictorNames})

Nonparametric

mannWhitney wilcoxon kruskalWallis friedman signTest runsTest kolmogorovSmirnovNormal

Reliability

cronbachAlpha(items, {names}) → ReliabilityResult

KMeans

cluster(points, {k, maxIter, seed, variableNames}) → KMeansResult

Distributions

normCdf normPdf normTwoTail tCdf tTwoTail tCritical
chiSquareCdf chiSquareSf chiSquareCritical
fCdf fSf fCritical betaInc lnGamma gammaFn
binomialPmf binomialSf combinations

Format

formatNum formatP significanceStars

测试

cd packages/statkit
dart pub get
dart test

当前覆盖：描述、分布、t 检验、ANOVA、相关、回归、非参数、信度、K-Means（13 组用例）。