files image

A enjoyable dinky venture to play with Jupyter Notebooks, Scikit-study, and neural nets with Keras.

Just

To speak a neural network to study Arabic morphology.

Involves:

  • Scripts for data mining
  • Starter data
  • 2 iterations of the model: roots.ipynb and roots-with-noroots.ipynb.

roots-with-noroots.ipynb is named unfamiliar, nevertheless it factual formula we are extra interesting about tracking words that are “mabniyy”, or undeclined.

It be no longer very accurate (about 50%) so or no longer it is beautiful addictive to work on. Completely, any person, someplace, has completed this better, nevertheless we’re no longer fixing world hunger here, factual having some nerdy enjoyable.

Pull requests welcome šŸ™‚

Pattern output

The output is formatted as “accuracy: [correctAnswer, input]” with output on the tip if unsuitable.

Right: ['??????', '???']
Ignored:  ['???????', '???'] Predicted: ???
Right: ['??????', '???']
...
Right: ['??????', '???']
Right: ['??????', '???']
Ignored:  ['????????', '???'] Predicted: ???
Ignored:  ['?????', '???'] Predicted: ???
Ignored:  ['???????', ''] Predicted: ???
Right: ['?????', '???']
Right: ['???', '???']
Right: ['????', '???']
Right: ['????', '???']
Ignored:  ['?????', '???'] Predicted: ???
Right: ['?????', '???']
Right: ['??????', '???']
Ignored:  ['???', '???'] Predicted: ???
Ranking: 58.8%

Easy the technique to use

Right install Jupyter Pocket e book and bustle jupyter pocket e book on this folder, and take grasp of one of the ipynb files.

Be taught Extra

Read More:  Starc injury 'can drag on a long time' - Cummins