data describe：
US Post Office Zip Code Data
Normalized handwritten digits, automatically scanned from envelopes by the U.S. Postal Service. The original scanned digits are binary and of different sizes and orientations; the images here have been deslanted and size normalized, resulting in 16 x 16 grayscale images (Le Cun et al., 1990).

The data are in two gzipped files, and each line consists of the digit id (0-9) followed by the 256 grayscale values.

There are 7291 training observations and 2007 test observations, distributed as follows:

    0     1     2     3     4     5     6     7     8     9     Total
Train     1194     1005     731     658     652     556     664     645     542     644     7291
Test     359     264     198     166     200     160     170     147     166     177     2007

or as proportions

    0     1     2     3     4     5     6     7     8     9
Train     0.16     0.14     0.1     0.09     0.09     0.08     0.09     0.09     0.07     0.09
Test     0.18     0.13     0.1     0.08     0.1     0.08     0.08     0.07     0.08     0.09

The test set is notoriously "difficult", and a 2.5% error rate is excellent. These data were kindly made available by the neural network group at AT&T research labs (thanks to Yann Le Cunn).