数据预处理
From Ufldl
Line 14: | Line 14: | ||
* 简单缩放 | * 简单缩放 | ||
- | * | + | * 逐样本均值消减(也称为移除直流分量) |
* 特征标准化(使数据集中所有特征都具有零均值和单位方差) | * 特征标准化(使数据集中所有特征都具有零均值和单位方差) | ||
Line 22: | Line 22: | ||
'''例子''':在处理自然图像时,我们获得的像素值在<math>[0, 255]</math>区间中,常用的处理是将这些像素值除以255,使它们缩放到<math>[0, 1]</math>中. | '''例子''':在处理自然图像时,我们获得的像素值在<math>[0, 255]</math>区间中,常用的处理是将这些像素值除以255,使它们缩放到<math>[0, 1]</math>中. | ||
- | === | + | === 逐样本均值消减 === |
【原文】 | 【原文】 | ||
If your data is ''stationary'' (i.e., the statistics for each data dimension follow the same distribution), then you might want to consider subtracting the mean-value for each example (computed per-example). | If your data is ''stationary'' (i.e., the statistics for each data dimension follow the same distribution), then you might want to consider subtracting the mean-value for each example (computed per-example). |