Usually the axis represent features and points represent the value. But for example in case of document classification, where each document is represented as a feature, what do the axis and points on the graph represent?
That's an example of an SVM classifier for 2-dimentional feature space. It means you've got only 2 real-valued features $x_1$ and $x_2$ and those circles and squares are objects. The painted objects are the support objects. You're right, for document classification you'd need to use more than 2 features, but it's more difficult to illustrate the concept in 3-dimentional space and impossible with higher dimensions. Also, for document classification you usually represent a document (an object) as a vector of features (not a single feature). These features could be tf-idf for instance.
$\begingroup$ In a 2d feature space diagram like the one from question, what do the axis represent? 2 words? Or 2 document classes? And what do the points represent? Words or documents? $\endgroup$– variableSep 10, 2020 at 13:22
$\begingroup$ 1) 2 words 2) 2 document classes are represented here with different color and shape of geometric figures: blue circles and red squares 3) points are documents $\endgroup$ Sep 10, 2020 at 15:02