reset password
Author Message
victormejia
Posts: 40
Posted 20:04 Feb 23, 2011 |

I get different t statistics every time I run my program, so sometimes the two accuracies are statistically significant, and sometimes not.

When I read in the training data, I randomly shuffle them, and I feed that data to both classifiers. Will you run our program a few times so that you can choose the best accuracy score?

cysun
Posts: 2935
Posted 07:37 Feb 24, 2011 |

That's a bit odd. I don't expect the accuracies to vary this much.

As for grading, I'll only run each program once.