Replace utility classes with sklearn classes in NB3#29
Conversation
| "T_KPCA_train = kpca.transform(X_train)\n", | ||
| "T_KPCA_test= kpca.transform(X_test)" |
There was a problem hiding this comment.
It's not that I forgot to deal with this comment, but I don't think using kernel_params = dict (kernel= "precomputed") is a good idea, and supplying kpca.transform with x_train is not an error that needs to be fixed immediately. I plan to deal with it after getting your reply.
| "outputs": [], | ||
| "source": [ | ||
| "krr = KRR(regularization=regularization, kernel_type=kernel_type)" | ||
| "krr = KernelRidge(alpha=regularization, kernel = 'rbf', gamma = 1.0)" |
| "outputs": [], | ||
| "source": [ | ||
| "krr.fit(X_train, Y_train, K=K_train)" | ||
| "krr.fit(X_train, Y_train)" |
There was a problem hiding this comment.
It's not that I forgot to deal with this comment, but I don't think using kernel_params = dict (kernel= "precomputed") is a good idea, and supplying krr.fit with x_train is not an error that needs to be fixed immediately. I plan to deal with it after getting your reply.
| "Y_krr_test = krr.transform(X_test, K=K_test)" | ||
| "Y_krr_train = krr.predict(X_train)\n", | ||
| "Y_krr_test = krr.predict(X_test)" | ||
| ] |
There was a problem hiding this comment.
What am I supposed to "see above"?
| " krr.statistics(X=X_test, Y=Y_test, K=K_test)], \n", | ||
| " headers = [\"Training\", \"Testing\"], \n", | ||
| " title=\"Kernel Ridge Regression\")" | ||
| "from utilities.classes import KPCovR # to be replaced with scalers from skcosmo" |
There was a problem hiding this comment.
It should be "to be replaced with the class from skcosmo"
PKGuo
changed the title
| "outputs": [], | ||
| "source": [ | ||
| "kpca.fit(X_train, K=K_train)" | ||
| "kpca.fit(X_train)" |
There was a problem hiding this comment.
then use kernel_params = dict(kernel="precomputed")!
|
Another reason for using pre-computed kernels…
Rose (Cersonsky) Adorf
(she/her)
rosecersonsky@gmail.com
… On Nov 20, 2020, at 4:18 PM, PKGuo ***@***.***> wrote:
@PKGuo commented on this pull request.
In notebooks/3_KernelMethods.ipynb <#29 (comment)>:
> "krr.fit(X_train, Y_train)\n",
- "Yhat_train = krr.transform(X_train).reshape(-1,Y.shape[-1])\n",
- "PKY = krr.PKY\n",
+ "Yhat_train = krr.predict(X_train).reshape(-1,Y.shape[-1])\n",
+ "PKY = np.linalg.solve(K_train + regularization*np.eye(n_train), \n",
Done. The problem is that KernelRidge won't automatically center the kernel matrix
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub <#29 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALKVP3S22DA3SOGFWJTWXZLSQ2CDBANCNFSM4TZVHM2Q>.
|
| "from utilities.kernels import linear_kernel, gaussian_kernel, center_kernel\n", | ||
| "from utilities.classes import PCA, LR, KPCA, KRR, KPCovR\n", | ||
| "\n", | ||
| "from utilities.kernels import linear_kernel, gaussian_kernel,center_kernel # to be replaced with functions from skcosmo\n", |
There was a problem hiding this comment.
if we don't use the kernel-tutorials kernels, please delete this line.
| "K_train_raw = kernel_func(X_train, X_train)" | ||
| "K_train_raw = kernel_func(X_train, X_train, gamma = 1.0)" |
There was a problem hiding this comment.
It will be centered later.
| "outputs": [], | ||
| "source": [ | ||
| "kpca.fit(X_train, K=K_train)" | ||
| "kpca.fit(X_train)" |
There was a problem hiding this comment.
it's not only about time, but about RAM. We have to assume that these notebooks will be used in a pedagogic setting, where students will be using computers of a variety of capabilities. It may be fast on your computer, but if we know it's a potential time-sink, we may as well avoid it.
| "T_KPCA_train = kpca.transform(X_train)\n", | ||
| "T_KPCA_test= kpca.transform(X_test)" |
| "outputs": [], | ||
| "source": [ | ||
| "krr.fit(X_train, Y_train, K=K_train)" | ||
| "krr.fit(X_train, Y_train)" |
| "Y_krr_test = krr.transform(X_test, K=K_test)" | ||
| "Y_krr_train = krr.predict(X_train)\n", | ||
| "Y_krr_test = krr.predict(X_test)" | ||
| ] |
| "# The string variable is used to pass to the utility classes. \n", | ||
| "kernel_func = gaussian_kernel\n", | ||
| "kernel_params = {\"kernel\": \"precomputed\"}\n", | ||
| "kernel_func = rbf_kernel\n", |
There was a problem hiding this comment.
Same comment as in #30, you can use functools to avoid having to include gamma everywhere
from functools import partial
kernel_func = partial(rbf_kernel, gamma=1.0)
| "Y_krr_test = krr.transform(X_test, K=K_test)" | ||
| "Y_krr_train = krr.predict(X_train)\n", | ||
| "Y_krr_test = krr.predict(X_test)" | ||
| ] |
There was a problem hiding this comment.
What am I supposed to "see above"?
|
Please double check that the notebook runs clean and then feel free to merge. |
2 participants
Replace some utility classes in
3_KernelMethods.ipynbwith appropriatesklearnclasses and modify the corresponding text.