linreg-core 0.8.1

Lightweight regression library (OLS, Ridge, Lasso, Elastic Net, WLS, LOESS, Polynomial) with 14 diagnostic tests, cross validation, and prediction intervals. Pure Rust - no external math dependencies. WASM, Python, FFI, and Excel XLL bindings.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

{
  "test_name": "Breusch-Godfrey Test (Python - statsmodels)",
  "dataset": "prostate",
  "formula": "lcavol ~ lweight + age + lbph + svi + lcp + gleason + pgg45 + lpsa",
  "order": 1,
  "statistic": 2.6663236458296353,
  "p_value": 0.10249252701586692,
  "test_type": "Chisq",
  "df": [
    1.0
  ],
  "f_statistic": 2.4590386609789383,
  "f_p_value": 0.1204804472155625,
  "f_df": [
    1.0,
    87.0
  ],
  "passed": true,
  "interpretation": "No significant serial correlation detected up to order 1",
  "description": "Tests for serial correlation up to order 1. The LM (Chi-squared) statistic is asymptotically distributed as chi-squared with 1 degrees of freedom. The F statistic provides a finite-sample correction. Uses statsmodels.stats.diagnostic.acorr_breusch_godfrey."
}