This page is about One-Hot-Encoding. One-Hot-Encoding can be used when you have sets of information either as inputs to a neural network or as outputs.

The Problem

Tensorflowjs works with real numbers (called decimal or floating point numbers), so if your Neural Network is outputing information that is about a defined group, the neural network should give a range from that group to the next group. Say you defined your output as 1.00 = mammals, 2=reptiles, 3=birds, etc.
Your neural network might identify an animal as 1.45. Meaning somewhere between Mammals and reptiles. So what does that mean?

The Solution

With One-Hot Encoding each output becomes an object to be rated by your neural network. So you might get results such as Mammal 75%, reptile 25%, bird 15%
This style of result becomes much more useful.

The Con

One-hot encoding needs large sets of inputs of zeros with only one input set as 1! This can quickly become a very large set of input data.

This Simple example

We will train the following information:

Species	Class	One-Hot-Species	One-Hot-Class
Elephant:	Mammal	1,0,0,0,0,0,0,0	1,0,0,0
Fox:	Mammal	0,1,0,0,0,0,0,0	1,0,0,0
Crow:	Bird (Aves)	0,0,1,0,0,0,0,0	0,1,0,0
Aligator:	Reptile	0,0,0,1,0,0,0,0	0,0,1,0
Dolphin:	Mammal	0,0,0,0,1,0,0,0	1,0,0,0
Sparrow:	Bird (Aves)	0,0,0,0,0,1,0,0	0,1,0,0
Turtle:	Reptile	0,0,0,0,0,0,1,0	0,0,1,0
Toad:	Amphibians	0,0,0,0,0,0,0,1	0,0,0,1

...

Now lets try the above in a full Machine Learning Web App

Tensorflowjs Machine Learning Web App Template

As simple as possible template. Note that for simplicity this has minimal CSS

Pre-Set Values:

Learning Rate:
Samples per batch:
Epochs: (# of batches to run)
Validation Split :

Data:

Elephant, Fox, Crow, Aligator, Dolphin, Sparrow, Turtle, Toad

Mammal, Mammal, Bird(aves), Reptile, Mammal, Bird (Aves), Reptile, Amphibian

xTrainingData:	1,0,0,0,0,0,0,0, 0,1,0,0,0,0,0,0, 0,0,1,0,0,0,0,0, 0,0,0,1,0,0,0,0, 0,0,0,0,1,0,0,0, 0,0,0,0,0,1,0,0, 0,0,0,0,0,0,1,0, 0,0,0,0,0,0,0,1
yTrainingTarget:	1,0,0,0, 1,0,0,0, 0,1,0,0, 0,0,1,0, 1,0,0,0, 0,1,0,0, 0,0,1,0, 0,0,0,1

Compile or Load Model:

Compile, Train, Test:			Import / Export
Compile Model:
Train Model:		Epochs run:
Test Trained Model with New Data:	New Data:	Crow0,0,1,0,0,0,0,0		Shape:

...