RatBrain

Neural Network simulator

Explore the world of AI

Train and run Neural Networks on your Palm!

RatBrain uses the classic Feed Forward Back Propagation algorithm to adjust connection strengths (weights). It can have up to eight variables on the Input layer, has three neurons in the Hidden layer and an Output neuron.

Neural Net example:

A lender might use a Neural Net to prescreen applicants, they would use client history to train the Net.

First determine the relevant parameters and scale them for inputs.

1. Age divided by 10 would produce 2 through 9 for the first input.

2. Marital status could be entered as married=9 single=5 divorced=0.

3. Education, years minus 6 would cover 6th grade to 3 years college.

4. Income divided by 10000=0 to 9.

5.. Residence could be 9=own 5=lease 0=other.

6. Debt to income could be scaled from 9=good to 0=bad.

7. Years on job (0 through 9).

8. Sex, 9=female 0=male.

Here's how the NNtrain file would look after entering training facts from past clients...