babao.models package

Subpackages

• babao.models.tree package
  • Submodules
  • babao.models.tree.extremaModel module
  • babao.models.tree.macdModel module
  • babao.models.tree.tendencyModel module
  • Module contents

Submodules

babao.models.modelBase module

Base class for any model

class babao.models.modelBase.ABCModel

Bases: abc.ABC

Base class for any model

dependencies_class

List of models or inputs needed by the current model

These should be class, not instances!

load()

Load the model from self.model_file

need_training

Specify if the current model need to be trained

plot(since)

Plot the model predictions from ´since´ timestamp

predict(since)

Return a dataframe of prediction starting from ´since´ timestamp

save()

Save the model to self.model_file

train(since)

Train the model with data starting from ´since´ timestamp

Return the score of model.

babao.models.modelBase.addLookbacks(df, look_back)

Add lookback(s) (shifted columns) to each df columns

babao.models.modelBase.addLookbacks3d(arr, look_back)

Add lookback(s) (shifted columns) to each df columns Reshape the features to be keras-proof (3d)

babao.models.modelBase.getVerbose() → int

Transform our verbose level to match keras one

babao.models.modelBase.reshape(arr)

Reshape the features to be keras-proof

babao.models.modelManager module

The idea here is to give a common interface to all the models so you can use these wrappers to call all of them at once.

babao.models.modelManager.plotModels(since)

Plot all models

babao.models.modelManager.predictModelsMaybeTrade(since)

Call predict on the root model, then eventually trade based on the preduction

babao.models.modelManager.trainModels(since)

Train all models and save the awesome result

babao.models.rootModel module

Root Model, base of the models tree

class babao.models.rootModel.RootModel

Bases: babao.models.modelBase.ABCModel

Root Model, base of the models tree

Not modeling much, but handle the call of the dependencies predictions

dependencies_class = [<class 'babao.models.tree.extremaModel.ExtremaModel'>]

load()

Load the model from self.model_file

need_training = False

plot(since)

Plot the model predictions from ´since´ timestamp

predict(since)

Call predict on the dependencies, then somehow merge the results

save()

Save the model to self.model_file

train(since)

Train the model with data starting from ´since´ timestamp

Return the score of model.

Module contents