Questions tagged [svm]
Support Vector Machine refers to "a set of related supervised learning methods that analyze data and recognize patterns, used for classification and regression analysis."
2,291 questions
5
votes
1
answer
127
views
Finding a linear classifier that optimizes AUC
A logistic regression will generate a set of hyperplanes like so:
(for more details about the picture see: Probability threshold in ROC curve analyses)
For computing the AUC of a classification the ...
- 96.3k
1
vote
0
answers
37
views
Support Vector Machine - Classification - General Advice
I wondered if it is possible to extract the 'support vectors' when doing a support vector machine (SVM) classifier?
For example, suppose I ran a 2-class SVM, can the two support vectors each side of ...
- 303
0
votes
0
answers
41
views
Is the figure showing margin violation for the support vector machine correct?
I am listening to a lecture on soft margin SVM https://youtu.be/XUj5JbQihlU?si=b66SblRnw9mmczVU&t=2969
The lecturer says that the blue dot represents a violation of the margin.
I don't really ...
1
vote
0
answers
42
views
Is the strong duality of the hard-margin SVM really trivially satisfied all the time?
It is widely known that if you were to calculate the maximizer of the dual SVM program (denote as $\alpha^*$), then the primal minimizer of the hard-margin SVM program,
\begin{aligned}&{\underset {...
2
votes
2
answers
180
views
Interpreting Shapley values for variance decomposition?
I trained a SVM multiple regression model and want to know how much each feature contributes to the prediction variance (quantified by the RMSE). I got the Shapley values for each feature on data from ...
- 209
0
votes
0
answers
36
views
Estimating the contribution of each feature to regression model prediction variance
I have trained a multiregression model using non-linear SVM, and got quite good metrics, with no big differences between test (20% data) and train (80% data) metrics.
The following are the test/train ...
- 209
2
votes
0
answers
55
views
What's the distribution of the lagrange multipliers found by quadratic programming? [closed]
I am trying to figure out how to infer C in support vector machine.
C is the upper bound on magnitude of lagrange multipliers. These multipliers are not independent. They are probably mutually ...
- 121
2
votes
1
answer
108
views
SVR model is learning rather too well
Memorability scores for a set of words can be downloaded from here. I am interested in seeing how well semantic embeddings can predict the relative memorability of words, as measured by Spearman's rho....
- 169
0
votes
0
answers
51
views
How to visualize Support Vector Regression in 3d
How would a visualization of the data underlying SVR, such as in
svm(X,y,type="nu-regression",kernel="linear",nu=0.5)
look like in 3 dimensions (...
- 757
0
votes
0
answers
108
views
Unrealistically high accuracy (100%) with ksvm model from kernlab in R
I'm trying to build an SVM model using the ksvm function from the kernlab package in R. My dataset is about breast cancer, and I'm trying to predict the diagnosis variable, which is a factor. All the ...
student95
1
vote
1
answer
81
views
Compare performance of support vector machine after feature deletion [closed]
I have two classes of data to train a linear support vector machine.
To be specific, I used Principle Component Analysis to project the data to 2-dimension and trained the support vector machine.
I ...
user398751
1
vote
0
answers
80
views
Best way to tackling SVM fine-tuning
I'm encountering a multiclass classification problem where I'm trying to predict 4 categories using SVM. I'm trying to fine-tuning its hyperparameter using Bayesian Optimization to speed up the ...
- 33
1
vote
1
answer
216
views
High Classification Accuracy Despite Poor Separation in PCA for Multi-class Data
I recently conducted a Principal Component Analysis (PCA) on a dataset with a four-category target variable. While the PCA score plot revealed excellent separation for one group, the remaining three ...
- 57
3
votes
1
answer
268
views
Why Mercer's Theorem is always cited for kernel learning and not Moore-Aronszajn
Why in most explanations of the kernel trick is Mercer's Theorem used as justification?
Can we not justify it as well with Moore-Aronszajn, which does not place the assumption of compactness on $X$ ...
- 117
2
votes
1
answer
133
views
Relation of RKHS to Feature Space in kernel learning
In standard SVM formulations, we typically look for a vector $w \in \mathbb{R}^D$ that defines a hyperplane in $\mathbb{R}^D$. The decision function is then of the form:
$$
f(x) = \operatorname{sign}(\...
- 117
