-
```
In subspace.cpp,around line 310:
double s = 1.00 / sqrt(eigenvalues[i][0]);
If the eigenvalue is zero or a negative number, the result will not be correct.
I am doing a PCA for sampletype img so…
-
PCA, SVD(其他low rank*), LDA(Topic Model), K-means, Sparse Coding,Hidden Layer of Neural Network。。。等等这一大类问题应该都可以用一套理论来解释其几何意义,与向量空间、矩阵、特征值和特征向量有关的,有阐述最本质原理的文章? 比特征值和特征向量更低层的?@好东西传送门
http://www.weibo.com…
-
-
When browsing linear regression, the hovered learning estimate for SVD disagreed with the value when clicking on the link, see
https://docs.google.com/file/d/0B5gSZqyvXnmXYm9kdXU1UEgybUE
cjrd updated
10 years ago
-
Occasionally I am annoyed by output from scikit-learn that munges words together due to multi-line string continuations in the code that forget to include whitespace.
Here's a collection of them to f…
-
This is the output
1.#QNAN, 1.#QNAN, 1.#QNAN, 1.#QNAN, 1.#QNAN, 1.#QNAN, 1.#QNAN, 1.#QNAN
1.#QNAN, 1.#QNAN, 1.#QNAN, 1.#QNAN, 1.#QNAN, 1.#QNAN, 1.#QNAN, 1.#QNAN
1.#QNAN, 1.#QNAN, 1.#QNAN, 1.#QNAN, 1.…
-
I start emacs with option -Q.
The new-loaded lusty-explorer works fine.
And then I set the frame font bigger (e.g. 16, DejaVu Sans Mono),
the problem appears.
Here is the debug backtrace:
Debugger e…