How to identify the corresponding index to target the categorical values as if there is 10 targets which one is for the largest number.

That’s covered in Week 2.

A sneak peek to week 2:

An example of categorical values for 3 targets (it will work the same for your 10 targets):


As you know, the computer will not understand Horse, Duck, Fish. So you will convert this to numbers:

0 - Horse
1 - Duck
2 - Fish

In the final vector where the classification happens, the position 0 will be for Horse, position 1 will be for Duck, and position 2 will be for fish.

Now, when the softmax processes the data, it will return this vector with 3 numbers that sum up to 1. For instance:

Position 0: 0.35
Position 1: 0.05
Position 2: 0.60

In this case, Position 2 has the highest probability, so the class that the algorithm is finding is #2, which is… “Fish”.

Hope this sheds light, and when you get to Week 2 you’ll be an advanced learner :slight_smile:



1 Like