数据预处理
From Ufldl
Line 1: | Line 1: | ||
- | |||
== 概要 == | == 概要 == | ||
Line 103: | Line 102: | ||
MNIST 数据集的像素值在 <math>[0, 255]</math> 区间中。我们首先将其缩放到 <math>[0, 1]</math> 区间。实际上,进行逐样本均值消去也有助于特征学习。''注:也可选择以对 MNIST 进行 PCA/ZCA 白化,但这在实践中不常用。'' | MNIST 数据集的像素值在 <math>[0, 255]</math> 区间中。我们首先将其缩放到 <math>[0, 1]</math> 区间。实际上,进行逐样本均值消去也有助于特征学习。''注:也可选择以对 MNIST 进行 PCA/ZCA 白化,但这在实践中不常用。'' | ||
+ | |||
+ | ==中英文对照== | ||
+ | |||
+ | :归一化 normalization | ||
+ | |||
+ | :白化 whitening | ||
+ | |||
+ | :直流分量 DC component | ||
+ | |||
+ | :局部均值消减 local mean subtraction | ||
+ | |||
+ | :消减归一化 sparse autoencoder | ||
+ | |||
+ | :缩放 rescaling | ||
+ | |||
+ | :逐样本均值消减 per-example mean subtraction | ||
+ | |||
+ | :特征标准化 feature standardization | ||
+ | |||
+ | :平稳 stationary | ||
+ | |||
+ | :Mel倒频系数 MFCC | ||
+ | |||
+ | :零均值化 zero-mean | ||
+ | |||
+ | :低通滤波 low-pass filtering | ||
+ | |||
+ | :基于重构的模型 reconstruction based models | ||
+ | |||
+ | :自编码器 autoencoders | ||
+ | |||
+ | :稀疏编码 sparse coding | ||
+ | |||
+ | :受限Boltzman机 RBMs | ||
+ | |||
+ | :k-均值 k-Means | ||
+ | |||
+ | :长尾 long tail | ||
+ | |||
+ | :损失函数 loss function | ||
+ | |||
+ | :正交化 orthogonalization | ||
==中文译者== | ==中文译者== | ||
- | @ | + | 陈磊(lei.chen@operasolutions.com), 王文中(wangwenzhong@ymail.com), 王方(fangkey@gmail.com) |
+ | |||
+ | |||
+ | {{Languages|Data_Preprocessing|English}} |