Visualization of Decision boundaries



Hi Guys,

Just sharing a good visualization of decision boundaries for various classifiers with different type of dataset, Hope you will find the image interesting.



Hey! Great visualization! Could you share the code which produced this?


Awesome work @aayushmnit. Can you please share the code ?


@jalFaizy @saurabh090909

Here you go. 7 - Comparison of Multiple Classification (519.4 KB)


@aayushmnit I m not able to download the zip file from the attached link. When clicked on link it say page can’t be found . Attached the screenshot . Please have a look.


@saurabh090909 Can you check now?


Thanks @kunal … Now its working :slight_smile:


So, how do you interpret these?


Well it’s to appreciate different algorithms , how they build their decision boundaries and to say that there is no one algorithm which can outperform every other algorithm.

If you are saying something like what do this mean. Then Red and blue labels are two classes on what we are fitting these algorithms and seeing the probability distribution determined by these algorithms with accuracy measure as a text label down right side.

~Request: Please be more specific with what you are trying to ask on portals. That will save some time for you and the person who’s answering.


@aayushmnit Thanks for the quick reply. I really meant to ask how to interpret the charts. I understand that the dots are our outcomes. The blue-red regions, I assume, show how the algorithm classifies the data. If this is so, what do the different shades mean? Are they probabilities? Dark colors meaning the classifier is more certain of the prediction?

Thanks for sharing.


@dgenchev : Yeah that’s exactly what it is. I have attached the code , why don’t you open and run for yourself?


Will do. Just didn’t have the time while at work, but I am looking forward to playing with it over the weekend.