This is the foundation of our reimplementation of the core ADWIN algorithm. It basically contains the checkHistogramForCut which receives a {@link His

The {@link ADWINInterface } is the basic interface for an external usage of this library.

Original implementation of ADWIN see https://github.com/abifet/adwin

The {@link ADWINWrapper} is used to combine a given {@link Histogram} with an given implementation of {@link ADWIN}

This is the implementation of the {@link ADWINInterface} for the original ADWIN implementation {@link ADWINOld}

The basic bucket is used in the histogram to contain the basic informations over elements.

This {@link BucketContainer} data structure is used in the Histogram and represents an array buckets.

The {@link DataGenerator} generates a stream of events. According to its parameters it generates different types of concept drifts. For example: ABRUP

This implementation of ADWIN contains the implementation of the HalfCutCheck approach. It uses a ExecuterService to create new Threads.

Datastructure for the exponential histogram used in ADWIN.

This is the benchmark for the evaluation of ADWIN. We Insert 100 batches with 1 mio. elements with a constant value This Micobenchmark makes use of ou

This is the serial implementation of ADWIN. It basically executes a full cut detection in the main thread.

This is an optimistic implementation of the ADWIN algorithm. It uses a separate thread for the adwin calculation, which is executed on a copy of the h

Basic abstraction to manage the different threads.