# Tag Info

0

You can use a simple lineplot.

0

Not sure how this is done in the any specific library, but here is what I would try. Given two sets of points ($K$ points in the first set and $S$ in the second set): $\{\mathbf{a}_i\}_{i=1\dots K}$ and $\{\mathbf{b}_i\}_{i=1\dots S}$, all in $N$-dimensional space ($\mathbb{R}^N$) I would compute the 'centres of mass' of the two sets:  \begin{align} \...

0

You have to add a constraint of linear separability to make the comparison. The problem is that in reality most problem are not linearly separable. Either they need non-linear transformations (data processing or usage of kernel), or they are not separable at all (underlying randomness, noise). SVM can be adapted to these non-linearities quite well (allowing ...

0

There are a couple potential things going on here. Your split is not homogeneus Your train and test are not similar, this can be done because you have split with a bias/pattern and both are different. This makes the train test error different. Overfitting Since the training error is better than the test, your model can just be overfitting. This kind of ...

1

Well, based on how PCA works, and the fact that you're using multi-class SVM, which is a pretty solid family of algorithm, the only possibility i can think of is that the problem comes from your data. It can still depends on the implementation you're using, but i think it come from your data.

1

I believe I have found the answer. A great thank-you to @Javier TG for pointing me to Andrew Ng's notes CS229 Suppor Vector Machines. I would not write down the full context as materials in the lecture notes and questions references do a much better job than I ever will. I will, however, try to fill in the tiny logic gap in many notes: why $w*x+b=1$ can ...

Top 50 recent answers are included