I have a question on the decision boundary for classification. In all the online tutorials, decision boundary are usually 2-d plot. For example, x vs y. However, I would like to draw a decision boundary between many variable, for example, x1,x2,x3,y1,y2,y3, would that be possible? For example, for FisherIris, would it be able to plot Sepal width, sepal length, pedal width and pedal length all together?
thank you in advance
When it comes to more than 2 dimensions, it becomes difficult because we have to visualize everything on a 2d plane (computer screen, paper etc.). 3d is still doable somehow, but with 4d and higher you have a problem. I think the best you can do is to make 2d plots of all feature combinations.
When it comes to more than 2 dimensions, it becomes difficult because we have to visualize everything on a 2d plane (computer screen, paper etc.). 3d is still doable somehow, but with 4d and higher you have a problem. I think the best you can do is to make 2d plots of all feature combinations.
Usually you combine variables in pairs, is difficult to visualize high dimensional data.
Bidimensional and tridimensional plots are the more used. My advice will be to select fixed critical points for one variable and to plot the other variables in such critical points. By doing this, you will obtain several plots (images) which you can play in a sequence to put together a video, providing an easy to understand representation.
Good Luck!
- Badges
- Science topic