This section contains software developed during my research. You are free to use any of the code provided for research purposes. If you have any questions or comments, let me know.

  • MOA extensions - a set of classes created during the writing of my master's thesis and my later research. Most of these classes are already part of the official MOA release.
  • Incremental modifications of block-based classifiers - source code of classes used to test incremental and online versions of the AWE and AUE1 algorithms, as described in "From Block-based Ensembles to Online Learners In Changing Data Streams: If- and How-To".
  • AUE2 dataset and plot scripts - scripts used to test the AUE2 algorithm, as described in "Reacting to Different Types of Concept Drift: The Accuracy Updated Ensemble Algorithm".
  • OAUE experiment scripts - scripts used to test the OAUE algorithm, as described in "Combining block-based and online methods in learning ensembles from concept drifting data streams".
  • XStreamClass - source code, datasets, and scripts used to test the XStreamClass algorithm, as described in "Adaptive XML Stream Classification using Partial Tree-edit Distance".
  • Prequential AUC - precompiled code, sources, datasets, and scripts used to test Prequential AUC and analyze its properties, as described in "Prequential AUC for Classifier Evaluation and Drift Detection in Evolving Data Streams" and "Prequential AUC: Properties of the Area Under the ROC Curve for Data Streams with Concept Drift".
  • Ensemble Diversity in Data Streams - visualizations of six different ensemble diversity measures calculated prequentially on streams with various types of concept drift. Precompiled code, sources, datasets, and scripts used to create these visualizations and use diversity measures for drift detection, as described in "Ensemble Diversity in Evolving Data Streams".
  • PO4 Coordination Clustering - datasets and python scripts used to cluster, visualize, and create new restraints from PO4 fragments, as described in "Conformation-dependent restraints for polynucleotides: I. Clustering of the geometry of the phosphodiester group".
  • Confirmation Measures in Rule-based Classification - CM-CAR algorithm source codes and supplementary materials, as described in "Bayesian Confirmation Measures in Rule-based Classification".
  • RestraintLib - webserver for generating phosphodiester structural restraints.
  • Using Network Analysis to Improve Nearest Neighbor Classification of Non-Network Data - R source codes and experiment scripts
  • Tetrahedron Measure Visualization - interactive web application tool that can aid the analysis of complete ranges of classifier performance and rule interestingness measures. R code available on Github.


Dariusz Brzeziński, Ph.D.
Institute of Computing Science
Poznan University of Technology
Piotrowo 2,
60-965 Poznan, Poland

BTiCW, room 2.7.13/5
(+48) 61 665-30-57
Tuesday 9:30-11:15