models.ElasticNetFit
A fitted elastic net, from ElasticNet.fit or ElasticNet.path.
Usage
models.ElasticNetFit()Attributes
| Name | Description |
|---|---|
| coefficients | Coefficients on the design’s scale; exact zeros where the penalty |
| deviance | Residual deviance. |
| deviance_ratio |
Share of the null deviance explained (glmnet’s dev.ratio).
|
| df | Number of non-zero coefficients, intercept excluded. |
| dispersion | Dispersion. |
| fitted | Fitted means on the training data. |
| input_hash | Hash of the training data (design, offset, weights, response). |
| lam | Penalty strength. |
| names | Coefficient names. |
| null_deviance | Deviance with only the intercept and unpenalized columns. |
coefficients
Coefficients on the design’s scale; exact zeros where the penalty
coefficients: list[float]
dropped a column.
deviance
Residual deviance.
deviance: float
deviance_ratio
Share of the null deviance explained (glmnet’s dev.ratio).
deviance_ratio: float
df
Number of non-zero coefficients, intercept excluded.
df: int
dispersion
Dispersion.
dispersion: float
fitted
Fitted means on the training data.
fitted: list[float]
input_hash
Hash of the training data (design, offset, weights, response).
input_hash: str
lam
Penalty strength.
lam: float
names
Coefficient names.
names: list[str]
null_deviance
Deviance with only the intercept and unpenalized columns.
null_deviance: float
Methods
| Name | Description |
|---|---|
| from_json() | Reads an artifact written by to_json. |
| predict() | Expected response for each row. |
| predict_distribution() | Joint predictive distribution across the rows, keyed |
| to_json() | The fit as a versioned JSON artifact (spec with its lambda and alpha, coefficients, fit statistics), with provenance (crate |
from_json()
Reads an artifact written by to_json.
Usage
from_json(text)Parameters
text: str
Returns
ElasticNetFit
Raises
ValueError- For malformed JSON, another format, a newer format version or inconsistent fields.
predict()
Expected response for each row.
Usage
predict(design)Parameters
design: Design- Same columns as the training design.
Returns
list of float
predict_distribution()
Joint predictive distribution across the rows, keyed
Usage
predict_distribution(design, n_sims, seed)row = 0, 1, ...: process uncertainty only (penalized coefficients have no standard errors; bootstrap the fit for parameter uncertainty).
Parameters
design: Designn_sims: intseed: int
Returns
PredictiveDistribution
to_json()
The fit as a versioned JSON artifact (spec with its lambda and alpha, coefficients, fit statistics), with provenance (crate
Usage
to_json()version and a hash of the training data). ElasticNetFit.from_json reads it back exactly; pickling uses it too.
Returns
str