如何读取R中的MNIST数据库？

endian="big",不是"high":

> to.read = file("~/Downloads/t10k-images-idx3-ubyte", "rb")

幻数:

> readBin(to.read, integer(), n=1, endian="big")
[1] 2051

图像数量:

> readBin(to.read, integer(), n=1, endian="big")
[1] 10000

行数:

> readBin(to.read, integer(), n=1, endian="big")
[1] 28

列数:

> readBin(to.read, integer(), n=1, endian="big")
[1] 28

这里有数据:

> readBin(to.read, integer(), n=1, endian="big")
[1] 0
> readBin(to.read, integer(), n=1, endian="big")
[1] 0

根据网站上的训练集图像数据描述.

现在你只需要将28*28字节块循环并读入矩阵.

重新开始:

 > to.read = file("~/Downloads/t10k-images-idx3-ubyte", "rb")

跳过标题:

> readBin(to.read, integer(), n=4, endian="big")
[1]  2051 10000    28    28

应该真正从标题读取28,28但硬编码在这里:

 > m = matrix(readBin(to.read,integer(), size=1, n=28*28, endian="big"),28,28)
 > image(m)

可能需要移调或翻转矩阵,我认为它是一个颠倒的"7".

par(mfrow=c(5,5))
par(mar=c(0,0,0,0))
for(i in 1:25){m = matrix(readBin(to.read,integer(), size=1, n=28*28, endian="big"),28,28);image(m[,28:1])}

得到你:

哦,谷歌引导我:http://www.inside-r.org/packages/cran/darch/docs/readMNIST这可能有用.