Letter dataset

This dataset was obtained from the UCI repository of machine learning databases. The task in this dataset is to guess the letter category (ie one out of 26 upper case printed letters) based on a 16-dimensional input vector. The inputs are simple global features computed from the images. More information is available in the original documentation associated with this dataset

Dataset profile:

Origin: simulated

Usage: development

Number of attributes: 17

Number of cases: 20,000

Number of prototasks: 1

Number of methods run on this dataset: 8

Download letter.tar.gz

Converted to Delve format by: Michael Revow

Last Updated 15 November 1996
Comments and questions to: delve@cs.toronto.edu