USPSdigits {IMIFA}

USPS handwritten digits

Description

Training and test sets for the United States Postal Service (USPS) handwritten digits data, with 8-bit 16x16 grayscale grid representations of image scans of the digits "0" through "9".

Usage

data(USPSdigits)

Format

A list of length 2 with the following elements, each one a data.frame:

train

The training set of 7,291 digits.

test

The test set of 2,007 digits.

Each data.frame contains the known digit labels in its first column.

The remaining 256 columns give the concatenation of the 16x16 grid.

Pixels are scaled such that [-1,1] corresponds to [white,black].

References

Hastie, T., Tibshirani, R., and Friedman, J. (2001). The Elements of Statistical Learning (2^nd edition). Springer Series in Statistics. New York, NY, USA: Springer.

See Also

show_digit, show_IMIFA_digit

Examples

# Load the data and record the labels
data(USPSdigits, package="IMIFA")
ylab  <- USPSdigits$train[,1]
train <- USPSdigits$train[,-1]

# Examine the effect of discarding peripheral pixels
SDs   <- apply(train, 2, sd)
ind   <- SDs > 0.7
dat   <- train[,ind]

hist(SDs, breaks=200, xlim=c(0, 1))
rect(0.7, 0, 1, 12, col=2, density=25)

show_digit(ind) # retained pixels are shown in black