Skip to main content

The stream-learn module is a set of tools necessary for processing data streams using scikit-learn estimators.

Project description


Travis Status Coverage Status CircleCI Status PyPI version

The stream-learn module is a set of tools necessary for processing data streams using scikit-learn estimators. The batch processing approach is used here, where the dataset is passed to the classifier in smaller, consecutive subsets called chunks. The module consists of five sub-modules:

  • streams - containing a data stream generator that allows obtaining both stationary and dynamic distributions in accordance with various types of concept drift (also in the field of a priori probability, i.e. dynamically unbalanced data) and a parser of the standard ARFF file format.
  • evaluators - containing classes for running experiments on stream data in accordance with the Test-Then-Train and Prequential methodology.
  • classifiers - containing sample stream classifiers,
  • ensembles - containing standard team models of stream data classification,
  • utils - containing typical classification quality metrics in data streams.

You can read more about each module in the documentation page.

Citation policy

If you use stream-learn in a scientific publication, we would appreciate citations to the following paper:

  title={stream-learn--open-source Python library for difficult data stream batch analysis},
  author={Ksieniewicz, Pawe{\l} and Zyblewski, Pawe{\l}},
  journal={arXiv preprint arXiv:2001.11077},

Quick start guide


To use the stream-learn package, it will be absolutely useful to install it. Fortunately, it is available in the PyPI repository, so you may install it using pip:

pip3 install -U stream-learn

You can also install the module cloned from Github using the file if you have a strange, but perhaps legitimate need:

git clone
cd stream-learn
make install

Preparing experiments

1. Classifier

In order to conduct experiments, a declaration of four elements is necessary. The first is the estimator, which must be compatible with the scikit-learn API and, in addition, implement the partial_fit() method, allowing you to re-fit the already built model. For example, we'll use the standard Gaussian Naive Bayes algorithm:

from sklearn.naive_bayes import GaussianNB
clf = GaussianNB()

2. Data Stream

The next element is the data stream that we aim to process. In the example we will use a synthetic stream consisting of shocking number of 100 chunks and containing precisely one concept drift. We will prepare it using the StreamGenerator() class of the stream-learn module:

from strlearn.streams import StreamGenerator
stream = StreamGenerator(n_chunks=100, n_drifts=1)

3. Metrics

The third requirement of the experiment is to specify the metrics used in the evaluation of the methods. In the example, we will use the accuracy metric available in scikit-learn and the precision from the stream-learn module:

from sklearn.metrics import accuracy_score
from strlearn.metrics import precision
metrics = [accuracy_score, precision]

4. Evaluator

The last necessary element of processing is the evaluator, i.e. the method of conducting the experiment. For example, we will choose the Test-Then-Train paradigm, described in more detail in User Guide. It is important to note, that we need to provide the metrics that we will use in processing at the point of initializing the evaluator. In the case of none metrics given, it will use default pair of accuracy and balanced accuracy scores:

from strlearn.evaluators import TestThenTrain
evaluator = TestThenTrain(metrics)

Processing and understanding results

Once all processing requirements have been met, we can proceed with the evaluation. To start processing, call the evaluator's process method, feeding it with the stream and classifier::

evaluator.process(stream, clf)

The results obtained are stored in the scores atribute of evaluator. If we print it on the screen, we may be able to observe that it is a three-dimensional numpy array with dimensions (1, 29, 2).

  • The first dimension is the index of a classifier submitted for processing. In the example above, we used only one model, but it is also possible to pass a tuple or list of classifiers that will be processed in parallel (See User Guide).
  • The second dimension specifies the instance of evaluation, which in the case of Test-Then-Train methodology directly means the index of the processed chunk.
  • The third dimension indicates the metric used in the processing.

Using this knowledge, we may finally try to illustrate the results of our simple experiment in the form of a plot::

import matplotlib.pyplot as plt


for m, metric in enumerate(metrics):
    plt.plot(evaluator.scores[0, :, m], label=metric.__name__)

plt.title("Basic example of stream processing")
plt.ylim(0, 1)


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for stream-learn, version
Filename, size File type Python version Upload date Hashes
Filename, size stream_learn- (57.7 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size stream-learn- (26.3 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page