Ensembles of Adaptive Model Rules from High-Speed Data Streams

João Duarte, João Gama
Proceedings of the 3rd International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications, PMLR 36:198-213, 2014.

Abstract

The volume and velocity of data is increasing at astonishing rates. In order to extract knowledge from this huge amount of information there is a need for efficient on-line learning algorithms. Rule-based algorithms produce models that are easy to understand and can be used almost offhand. Ensemble methods combine several predicting models to improve the quality of prediction. In this paper, a new on-line ensemble method that combines a set of rule-based models is proposed to solve regression problems from data streams. Experimental results using synthetic and real time-evolving data streams show the proposed method significantly improves the performance of the single rule-based learner, and outperforms two state-of-the-art regression algorithms for data streams.

Cite this Paper

BibTeX
@InProceedings{pmlr-v36-duarte14, title = {Ensembles of Adaptive Model Rules from High-Speed Data Streams}, author = {Duarte, João and Gama, João}, booktitle = {Proceedings of the 3rd International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications}, pages = {198--213}, year = {2014}, editor = {Fan, Wei and Bifet, Albert and Yang, Qiang and Yu, Philip S.}, volume = {36}, series = {Proceedings of Machine Learning Research}, address = {New York, New York, USA}, month = {24 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v36/duarte14.pdf}, url = {https://proceedings.mlr.press/v36/duarte14.html}, abstract = {The volume and velocity of data is increasing at astonishing rates. In order to extract knowledge from this huge amount of information there is a need for efficient on-line learning algorithms. Rule-based algorithms produce models that are easy to understand and can be used almost offhand. Ensemble methods combine several predicting models to improve the quality of prediction. In this paper, a new on-line ensemble method that combines a set of rule-based models is proposed to solve regression problems from data streams. Experimental results using synthetic and real time-evolving data streams show the proposed method significantly improves the performance of the single rule-based learner, and outperforms two state-of-the-art regression algorithms for data streams.} }
Endnote
%0 Conference Paper %T Ensembles of Adaptive Model Rules from High-Speed Data Streams %A João Duarte %A João Gama %B Proceedings of the 3rd International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications %C Proceedings of Machine Learning Research %D 2014 %E Wei Fan %E Albert Bifet %E Qiang Yang %E Philip S. Yu %F pmlr-v36-duarte14 %I PMLR %P 198--213 %U https://proceedings.mlr.press/v36/duarte14.html %V 36 %X The volume and velocity of data is increasing at astonishing rates. In order to extract knowledge from this huge amount of information there is a need for efficient on-line learning algorithms. Rule-based algorithms produce models that are easy to understand and can be used almost offhand. Ensemble methods combine several predicting models to improve the quality of prediction. In this paper, a new on-line ensemble method that combines a set of rule-based models is proposed to solve regression problems from data streams. Experimental results using synthetic and real time-evolving data streams show the proposed method significantly improves the performance of the single rule-based learner, and outperforms two state-of-the-art regression algorithms for data streams.
RIS
TY - CPAPER TI - Ensembles of Adaptive Model Rules from High-Speed Data Streams AU - João Duarte AU - João Gama BT - Proceedings of the 3rd International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications DA - 2014/08/13 ED - Wei Fan ED - Albert Bifet ED - Qiang Yang ED - Philip S. Yu ID - pmlr-v36-duarte14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 36 SP - 198 EP - 213 L1 - http://proceedings.mlr.press/v36/duarte14.pdf UR - https://proceedings.mlr.press/v36/duarte14.html AB - The volume and velocity of data is increasing at astonishing rates. In order to extract knowledge from this huge amount of information there is a need for efficient on-line learning algorithms. Rule-based algorithms produce models that are easy to understand and can be used almost offhand. Ensemble methods combine several predicting models to improve the quality of prediction. In this paper, a new on-line ensemble method that combines a set of rule-based models is proposed to solve regression problems from data streams. Experimental results using synthetic and real time-evolving data streams show the proposed method significantly improves the performance of the single rule-based learner, and outperforms two state-of-the-art regression algorithms for data streams. ER -
APA
Duarte, J. & Gama, J.. (2014). Ensembles of Adaptive Model Rules from High-Speed Data Streams. Proceedings of the 3rd International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications, in Proceedings of Machine Learning Research 36:198-213 Available from https://proceedings.mlr.press/v36/duarte14.html.

Related Material