Skip to main content

transform​

transform(
    @object ## llll (required)
    @x ## llll/list (required)
    @inverse 0
) -> llll

Transforms a dataset or point given some ML object such as a scaler (e.g., normscaler, stdscaler, robustscaler) or dimensionality reducer (e.g., umap, pca).

Arguments​

  • @object [llll]: Model object (required)
  • @x [llll/list]: Dataset or point to transform. (required)
  • @inverse [int]: Inverse transform mode. Ignored with @object is a UMAP (default: 0).
    • 0: Off
    • 1: On

Output​

Transformed dataset or data point [llll]

Usage​

$indataset = dataset(
    for $i in 1...100 collect [$i * 10 ** (-1...1)] ## dummy input dataset
);
$inpoint = 0.65 6.5 65; ## dummy input point
$scaler = normscaler($indataset); ## create scaler based on dummy dataset
$outdataset = transform($scaler, $indataset); ## transform dataset based on learned parameters
$outpoint = transform($scaler, $inpoint); ## transform new point based on learned parameters
writeobject($scaler, './normscaler.json'); ## write to JSON (optional)
$scaler = readobject('./normscaler.json'); ## read from JSON (optional)
print(getitems($indataset, 1...5), 'Input dataset:');
print(getitems($outdataset, 1...5), 'Output dataset:');
print($outpoint, 'Output point:')