0
$\begingroup$

I have implemented a support vector machine (SVM) in Python. I want to know the relationship between the number of features and the number of dimensions. My dataset contains 5 features, does it mean that my SVM classifier has 5 dimensions? Thanks

$\endgroup$
0
$\begingroup$

Excluding the target variable, if your dataset has $n$ columns (predictors, covariates etc.), then your SVM separates the space $\mathbb{R}^n$. So, Your SVM operates on $n$ dimensional data. The separating hyperplane is in $\mathbb{R}^{n-1}$, i.e. $n-1$ dimensional by definition, and its equation can be generally written as $w^Tx+b$, where $w,x$ are $n$ dimensional, and $b$, bias, is $1$ dimensional (i.e. a scalar). So, the hyperplane equation has $n+1$ coefficients, including the bias term, which means your parameter space is $n+1$ dimensional. I haven't heard a term like SVM dimension, but these are all the dimensionality concerns inside the classifier.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.