optdigits {mlr3data}

Optical Recognition of Handwritten Digits

Description

Classification data to predict handwritten digits. Obtained using the mlr3oml package.

Binarized version of the original data set. The multi-class target column has been converted to a two-class nominal target column by re-labeling the majority class as positive ("P") and all others as negative ("N"). Originally converted by Quan Sun.

Contains 64 features and 5620 observations. Target column is "binaryclass".

Pre-processing

• All feature variables "input1", ..., "input64" (number of on pixels in each block) have been coerced to integers.

• The target variable has been renamed from "binaryClass" to "binaryclass".

Source

https://www.openml.org/d/980

Examples

data("optdigits", package = "mlr3data")
str(optdigits)