search

devYonz / epsilon

Stanford ML Project CS229'18
1 stars 1 forks source link

Investigate feature engineering methods. #3

Open gksb88 opened 5 years ago

gksb88 commented 5 years ago
  1. Standard NLTK stuff like lematize, stop words
  2. BOW with n-gram or n-gram skip models
  3. word2vec
devYonz commented 5 years ago

I will start on this on Monday with a choice of three to experiment and apply to NB and SVM solutions.

gksb88 commented 5 years ago

Sounds good. I think thats where the problem lies. In the meanwhile, ill try to look at methods to improve accuracy on the neural net front.

devYonz commented 5 years ago

For tonight can you try saving the captures while training the net. Instead of showing the Image, basically the section for selecting an optimal neural network architecture

On Sun, Dec 2, 2018, 2:00 PM gksb88 notifications@github.com wrote:

Sounds good. I think thats where the problem lies. In the meanwhile, ill try to look at methods to improve accuracy on the neural net front.

— You are receiving this because you were assigned.

Reply to this email directly, view it on GitHub https://github.com/yfeleke/epsilon/issues/3#issuecomment-443546242, or mute the thread https://github.com/notifications/unsubscribe-auth/AIO_6GMwdpCpsNe7ZWzIuYDyWJQuP5nNks5u1E2AgaJpZM4YoXjW .

gksb88 commented 5 years ago

Will do