1.3.7 The explicit route🔗ℹ

Every procedure converts its data to a design matrix and a response on each call, and checks it each time. The conversions can also be called directly: to check the data once and fit it many times, to see the matrix a fit reads, or to convert a design matrix back to a format. Rows, columns and tables convert with glmnet itself (see Input data); each other format has its module:

A fit from the converted data is the fit from the data itself:

> (require glmnet/data/polars)

> (define measures     (polars->design-matrix flowers '("Sepal.Width" "Petal.Length" "Petal.Width")))

> (define sepal-length (polars->response flowers "Sepal.Length"))

> (equal? (lasso measures sepal-length #:lambda 0.01)           (lasso flowers "Sepal.Length" #:lambda 0.01                  #:predictors '("Sepal.Width" "Petal.Length" "Petal.Width")))

#t

> (for/list ([lam '(1.0 0.1 0.01)])     (coef (lasso measures sepal-length #:lambda lam)))

'(#(5.843333333333333 0.0 0.0 0.0)   #(3.85433085816735 0.19273279296149542 0.3724731460400136 0.0)   #(2.291509952135093     0.5717696868678742     0.534050705108672     -0.16945221616247616))

Column names are kept as strings, whether they were given as strings or symbols, so the same data named either way gives equal? design matrices. The columns of a design matrix with no names are named "V1", "V2" and so on, as R names them:

> (require glmnet/data/nested)

> (define D (nested->design-matrix rows #:column-names '(x1 x2)))

> (design-matrix-column-names D)

'("x1" "x2")

> (design-matrix->nested D #:by 'columns #:outer 'vector #:inner 'vector)

'#(#(1.0 2.0 3.0 4.0 5.0) #(2.0 1.0 4.0 3.0 6.0))

> (design-matrix->rows (table->design-matrix patients '("age" "dose")))

'((34.0 2.5) (51.0 5.0) (67.0 1.0) (45.0 3.5))

> (design-matrix->polars (design-matrix-select-rows measures '(0 50 100)))

shape: (3, 3) ┌─────────────┬──────────────┬─────────────┐ │ Sepal.Width ┆ Petal.Length ┆ Petal.Width │ │ ---         ┆ ---          ┆ ---         │ │ f64         ┆ f64          ┆ f64         │ ╞═════════════╪══════════════╪═════════════╡ │ 3.5         ┆ 1.4          ┆ 0.2         │ │ 3.2         ┆ 4.7          ┆ 1.4         │ │ 3.3         ┆ 6.0          ┆ 2.5         │ └─────────────┴──────────────┴─────────────┘

> (design-matrix->polars (rows->design-matrix '((1 2) (3 4))))

shape: (2, 2) ┌─────┬─────┐ │ V1  ┆ V2  │ │ --- ┆ --- │ │ f64 ┆ f64 │ ╞═════╪═════╡ │ 1.0 ┆ 2.0 │ │ 3.0 ┆ 4.0 │ └─────┴─────┘

> (table->polars (list (cons "y" '(1.5 2.5)) (cons "group" '("a" "b"))))

shape: (2, 2) ┌─────┬───────┐ │ y   ┆ group │ │ --- ┆ ---   │ │ f64 ┆ str   │ ╞═════╪═══════╡ │ 1.5 ┆ a     │ │ 2.5 ┆ b     │ └─────┴───────┘

design-matrix->polars, the way back, is slower than the way in.