brglm2 (version 1.0.1)

MultipleFeatures: Multiple features data

Description

Digits (0-9) extracted from a collection of maps from a Dutch public utility. Two hundred 30 × 48 binary images per digit are available, which have then been used to extract feature sets; see Jain et al. (2000), for details, where that dataset is used for assessing the performance of various classifiers for digit recognition.

Usage

MultipleFeatures

Format

A data frame with 2000 rows and 382 columns:

• digit. The digits to which the feature sets fou.*, kar.* and pix.* correspond to.

• fou.*. 76 Fourier coefficients of the character shapes, which are computed to be rotation invariant.

• kar.*. 64 Karhunen-Lo\`eve coefficients of the character shapes.

• pix.*. 240 pixel averages in 2 x 3 windows of each character shape.

• training. TRUE if the digit is part of the training set and FALSE if the digit is allocated to the test set.

References

Duin, R. (1998). Multiple Features Dataset. UCI Machine Learning Repository. tools:::Rd_expr_doi("10.24432/C5HC70").

Jain A, Duin R, Mao J (2000). Statistical pattern recognition: a review. IEEE Transactions on Pattern Analysis and Machine Intelligence, 22, 4–37. tools:::Rd_expr_doi("10.1109/34.824819").

See Also

mdypl_fit()

Examples

# \donttest{
data("MultipleFeatures", package = "brglm2")

par(mfrow = c(10, 20), mar = numeric(4) + 0.1)
for (c_digit in 0:9) {
   df <- subset(MultipleFeatures, digit == c_digit)
   df <- as.matrix(df[, paste("pix", 1:240, sep = ".")])
   for (inst in 1:20) {
       m <- matrix(df[inst, ], 15, 16)[, 16:1]
       image(m, col = grey.colors(7, 1, 0), xaxt = "n", yaxt = "n")
  }
}

# }