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
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36

{
  "test_name": "Cook's Distance (Python - statsmodels)",
  "dataset": "longley",
  "formula": "GNP.deflator ~ GNP + Unemployed + Armed Forces + Population + Year + Employed",
  "distances": [
    0.06486411793008785,
    0.24984688172798683,
    0.0009588205127364854,
    0.23278249660091008,
    0.1047580637084148,
    0.010083981691623885,
    0.12434139250895876,
    0.0030589764963460065,
    0.07881341713944455,
    0.06845889948340325,
    0.05830654678383011,
    0.13866534892149354,
    0.0155985011300882,
    0.0028808338221085424,
    0.008303997334998736,
    0.3402165604542402
  ],
  "p": 7,
  "mse": 1.4271113668658895,
  "threshold_4_over_n": 0.25,
  "threshold_4_over_df": 0.4444444444444444,
  "threshold_1": 1.0,
  "influential_4_over_n": [
    16
  ],
  "influential_4_over_df": [],
  "influential_1": [],
  "max_distance": 0.3402165604542402,
  "max_index": 16,
  "description": "Measures influence of each observation on regression coefficients. Uses statsmodels.stats.outliers_influence.OLSInfluence.cooks_distance."
}