mlbend-report

Releases (1)

• 0.1.0.0 indexed

Definitions (27)

main.bend

• type Cls
• type Report
• def cell.goThe cell of the true label i and the predicted label j: how many samples have both.
• def cell
• def rows_ofHow many rows a matrix has.
• def cells_ofHow many cells a row has.
• def row_fromThe row of the class i, from the column j on: n cells, the cell of every class from j to j+n-1, ascending.
• def rowThe row of the class i: k cells, ascending.
• def matrix_fromThe rows of the classes i to i+n-1: n rows, each of k cells, ascending.
• def matrixThe confusion matrix of k classes: k rows of k cells, the row of the class 0 first.
• law matrix_rowsLAW 1 (matrix_rows): n rows of the matrix are n rows.
• law row_lengthLAW 2 (row_length): n cells of a row are n cells.
• def cls_ofThe report of one class: its precision, its recall, its F1 and its support.
• def classes_of.goThe report of every class, from the class i on: n of them, ascending.
• def classes_ofThe report of every class: k of them, the class 0 first.
• type TotThe sums the averages rest on: the per-class scores added up, and the support beside them.
• def tot_addOne class more, added into the sums.
• def tot_of.goThe sums over the classes.
• def tot_of
• def dividea divided by b, or 0.0 when b is zero — the zero_division convention
• def macro_of.go
• def macro_ofThe macro average: the mean of the per-class scores, every class weighing the same, the way scikit-learn's macro average does.
• def weighted_sumsThe weighted sums: every per-class score multiplied by its support.
• def weighted_of.go
• def weighted_ofThe weighted average: the per-class scores, each weighed by its support, over the total support, the way scikit-learn's weighted average does.
• def report.goThe whole report, or why there is none: the labels and the predictions must agree in length, and there must be at least one sample.
• def report