Dataset - GateNLP/gate-lf-python-data GitHub Wiki

The Dataset class is used to represent training data. An instance of this class gets created by pointing to the meta file of the json file pair. After the instance has been created, the dataset instance contains the metadata in the .meta attribute.

After this the method instances_as_data can be used to iterate over the data returning lists of converted data: each list contains two sublists, the converted independent data and the converted dependent data.

The independent data is the result of converting the independent part of the json list using a features instance (see Features) through features(indep), similarly, the dependent data is the result of target(dep).