Results
- With the FLIF net, we got 52.81% averaged over 100 10-fold tests.
- We also did this with a Kohonen SOM and got 56.00%.
- In the past we did this with several other algorithms as a 10-fold
test (which is why we picked 10-fold).
- The best result was 57% using an MLP with backperc.
- We didn't spend a lot of time optimising performance, but
the biological system is in the ball park on this task.
