reset password
Author Message
306264301
Posts: 28
Posted 12:04 Apr 10, 2020 |

I have a question for part1. for splitting the data into train and test, do we use ?

test_size=0.2    and random_state=2

 

mpourhoma
Posts: 39
Posted 14:14 Apr 10, 2020 |

For Problem1, you don't split the data into testing and trainings manually. You have to Use 10-fold Cross-Validation to evaluate your model.

306264301
Posts: 28
Posted 14:38 Apr 10, 2020 |

thanks.

hassan87
Posts: 1
Posted 21:39 Apr 10, 2020 |
 
 

 

For Problem2) why c=1, Use SVC(C=1, kernel='rbf', gamma=0.0005) 

 
Last edited by hassan87 at 02:16 Apr 11, 2020.
mpourhoma
Posts: 39
Posted 02:15 Apr 11, 2020 |

In part f, you only try C=1. In part g, you do grid search to find the best C.