This video and your others have been very useful but I have a quick question:
My data set contains 4 different skull measurements from 2 different epochs. How do I account for the division in the dataset by epoch? I'm thinking that it would involve plotting the observations in colors in the biplot, each epoch being a different color. Is there a way to do that easily in R or am I better off taking it to Sigmaplot?
@WolferaTheJackal The predict function should work. Probably the reason you are getting really big values it that you are using the cor=TRUE which centers and scales the data. You would need to center and scale your Xnew data then project it.
@WolferaTheJackal Why don't you just use the scores=TRUE statement? That will project it into the space.
In your code the problem is that you are using *. This isn't matrix multiplication in R. It is elementwise multiplication. Matrix mutliplication is %*%. So you code should look like:
As I don't do a whole lot of work with SVMs from when I have used them I have used the libsvm package. Here is a link to a paper that has examples of using SVMs in a few different packages and an assessment of them. It also gives example code that you can run to get things going.
Journal of statistical software... version 15, issue 9.
The point of these videos is so that you don' t have to buy a book. They are here free. However the fact that you would buy a book from me is very flattering. Feel free to request other topics. I really don't know what people want to learn outside of basic statistical methods.
Thank you so much for your videos! You're doing great work indeed!
But would you be so kind to upload a brief example of using 'k-opls'? I'm really stuck with it... It seems quite difficult for a person without significant math background... I would be more than happy if you did that. And considering that this method is quite actual today and becoming more and more popular (in particular, in biomed science), it might be interesting for many people...
thank you. It is very useful
Khibanmaitrolai 3 weeks ago
This has been flagged as spam show
May I ask, why have you considered the post about kopls as a spam??
AimaNoima 2 months ago
This video and your others have been very useful but I have a quick question:
My data set contains 4 different skull measurements from 2 different epochs. How do I account for the division in the dataset by epoch? I'm thinking that it would involve plotting the observations in colors in the biplot, each epoch being a different color. Is there a way to do that easily in R or am I better off taking it to Sigmaplot?
DekeBlue 5 months ago
@WolferaTheJackal The predict function should work. Probably the reason you are getting really big values it that you are using the cor=TRUE which centers and scales the data. You would need to center and scale your Xnew data then project it.
Hope this helps.
off2themovies2 10 months ago
@WolferaTheJackal Why don't you just use the scores=TRUE statement? That will project it into the space.
In your code the problem is that you are using *. This isn't matrix multiplication in R. It is elementwise multiplication. Matrix mutliplication is %*%. So you code should look like:
Tnew <- Xnew%*%P
That should work.
off2themovies2 10 months ago
As I don't do a whole lot of work with SVMs from when I have used them I have used the libsvm package. Here is a link to a paper that has examples of using SVMs in a few different packages and an assessment of them. It also gives example code that you can run to get things going.
Journal of statistical software... version 15, issue 9.
off2themovies2 10 months ago 4
I think how support vector machines work in R depends on the SVM package you are using. Some packages are easier to use than others.
off2themovies2 10 months ago
The point of these videos is so that you don' t have to buy a book. They are here free. However the fact that you would buy a book from me is very flattering. Feel free to request other topics. I really don't know what people want to learn outside of basic statistical methods.
off2themovies2 11 months ago
Comment removed
AllphenX 2 months ago
This has been flagged as spam show
@off2themovies2
Thank you so much for your videos! You're doing great work indeed!
But would you be so kind to upload a brief example of using 'k-opls'? I'm really stuck with it... It seems quite difficult for a person without significant math background... I would be more than happy if you did that. And considering that this method is quite actual today and becoming more and more popular (in particular, in biomed science), it might be interesting for many people...
AllphenX 2 months ago 6
Comment removed
AimaNoima 2 months ago
I love your video, helped me so much~~~you should write a book....and I will buy~haha
insomnialu 11 months ago