A collection of supervised learning models based on shallow neural network approaches (e.g., word2vec and fastText) with some additional exclusive features. Written in Python and fully compatible with scikit-learn.
Install the latest version:
pip install cython
pip install shallowlearnImport models from shallowlearn.models, they implement the standard methods for supervised learning in scikit-learn,
e.g., fit(X, y), predict(X), etc.
Data is raw text, each sample is a list of tokens (words of a document), while each target value in y can be a
single label (or a list in case of multi-label training set) associated with the relative sample.
A supervised learning model based on the fastText algorithm [1]. The code is mostly taken and rewritten from Gensim, it takes advantage of its optimizations (e.g. Cython) and support.
It is possible to choose the Softmax loss function (default) or one of its two "approximations":
Hierarchical Softmax and Negative Sampling. It is also possible to load pre-trained word vectors at initialization,
passing a Gensim Word2Vec or a ShallowLearn LabeledWord2Vec instance (the latter is retrievable from a
GensimFastText model by the attribute classifier).
Constructor argument names are a mix between the ones of Gensim and the ones of fastText (see this class docstring).
>>> from shallowlearn.models import GensimFastText
>>> clf = GensimFastText(size=100, min_count=0, loss='hs', max_iter=3, random_state=66)
>>> clf.fit([('i', 'am', 'tall'), ('you', 'are', 'fat')], ['yes', 'no'])
>>> clf.predict([('tall', 'am', 'i')])
['yes']The supervised algorithm of fastText implemented in fastText.py ,
which exposes an interface on the original C++ code.
The current advantages of this class over GensimFastText are the subwords ant the n-gram features implemented
via the hashing trick.
The constructor arguments are equivalent to the original supervised model, except for input_file, output and
label_prefix.
WARNING: The only way of loading datasets in fastText.py is through the filesystem (as of version 0.8.0),
so data passed to fit(X, y) will be written in temporary files on disk.
>>> from shallowlearn.models import FastText
>>> clf = FastText(dim=100, min_count=0, loss='hs', epoch=3, bucket=5, word_ngrams=2)
>>> clf.fit([('i', 'am', 'tall'), ('you', 'are', 'fat')], ['yes', 'no'])
>>> clf.predict([('tall', 'am', 'i')])
['yes']TODO: Based on https://radimrehurek.com/gensim/models/word2vec.html#gensim.models.word2vec.Word2Vec.score
TODO: future features are going to be listed as Issues
The script scripts/document_classification_20newsgroups.py refers to this
scikit-learn example
in which text classifiers are compared on a reference dataset;
we added our models to the comparison.
The current results, even if still preliminary, are comparable with other
approaches, achieving the best performance in speed.
Results as of release 0.0.3,
with chi2_select option set to 80%.
The times take into account of tf-idf vectorization in the “classic” classifiers, and the I/O operations for the
training of fastText.py.
Test time of GensimFastText is the sum of the single predictions for each test sample,
it must be optimized in a unique, monolithic prediction.
The evaluation measure is macro F1.
| [1] |
|