PCA Leading to Increased Training Time

[ageron]

PCA Leading to Increased Training Time

[ageron]

Hi @ndalton12 ,
As admiral Ackbar would say, it's a trap! ;-)
Reducing dimensionality sometimes reduces training time significantly, but in some cases it can actually lead to the reverse. It depends on the model, the dataset, the training algorithm and possibly the weather (check out Figure 8-6). If you try a software regression classifier (using Scikit-Learn's LogisticRegression class), you will see that training gets a ×4 speedup on MNIST when using PCA.
I just added the solution to exercise 9 if you want to take a look.
Thanks for your kind words, I'm glad you are enjoying the book.
Cheers,
Aurélien

[Mauricio9]

???? mandou por engano

…

On Wed, May 10, 2023 at 4:06 PM defersha ***@***.***> wrote: Thanks. I was facing the same issue! — Reply to this email directly, view it on GitHub <#44 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACQPMJQNBGIA6C47STWV42TXFPRL5ANCNFSM4DQQXGNQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>