Communications of the ACM
Mathematical statistics (4th ed.)
Mathematical statistics (4th ed.)
Acquisition of dynamic control knowledge for a robotic manipulator
Proceedings of the seventh international conference (1990) on Machine learning
Instance-Based Learning Algorithms
Machine Learning
Tracking drifting concepts using random examples
COLT '91 Proceedings of the fourth annual workshop on Computational learning theory
Learning time-varying concepts
NIPS-3 Proceedings of the 1990 conference on Advances in neural information processing systems 3
Learning flexible concepts from streams of examples: FLORA2
ECAI '92 Proceedings of the 10th European conference on Artificial intelligence
Learning and forgetting for perception-action: a projection pursuit and density adaptive approach
Learning and forgetting for perception-action: a projection pursuit and density adaptive approach
An Algorithm for Finding Best Matches in Logarithmic Expected Time
ACM Transactions on Mathematical Software (TOMS)
Incremental Learning from Noisy Data
Machine Learning
Pattern Classification (2nd Edition)
Pattern Classification (2nd Edition)
When Experience Is Wrong: Examining CBR for Changing Tasks and Environments
ICCBR '99 Proceedings of the Third International Conference on Case-Based Reasoning and Development
Reactive and Memory-Based Genetic Programming for Robot Control
Proceedings of the Second European Workshop on Genetic Programming
Incremental rule learning based on example nearness from numerical data streams
Proceedings of the 2005 ACM symposium on Applied computing
Combining proactive and reactive predictions for data streams
Proceedings of the eleventh ACM SIGKDD international conference on Knowledge discovery in data mining
Data streams classification by incremental rule learning with parameterized generalization
Proceedings of the 2006 ACM symposium on Applied computing
PointMap: A Real-Time Memory-Based Learning System with On-line and Post-Training Pruning
International Journal of Hybrid Intelligent Systems
Efficient instance-based learning on data streams
Intelligent Data Analysis
Catching the Drift: Using Feature-Free Case-Based Reasoning for Spam Filtering
ICCBR '07 Proceedings of the 7th international conference on Case-Based Reasoning: Case-Based Reasoning Research and Development
Mining Concept-Drifting Data Streams with Multiple Semi-Random Decision Trees
ADMA '08 Proceedings of the 4th international conference on Advanced Data Mining and Applications
A case-based technique for tracking concept drift in spam filtering
Knowledge-Based Systems
A new learning strategy for classification problems with different training and test distributions
IWANN'07 Proceedings of the 9th international work conference on Artificial neural networks
Quick adaptation to changing concepts by sensitive detection
IEA/AIE'07 Proceedings of the 20th international conference on Industrial, engineering, and other applications of applied intelligent systems
An efficient algorithm for instance-based learning on data streams
ICDM'07 Proceedings of the 7th industrial conference on Advances in data mining: theoretical aspects and applications
Classification model for data streams based on similarity
IEA/AIE'11 Proceedings of the 24th international conference on Industrial engineering and other applications of applied intelligent systems conference on Modern approaches in applied intelligence - Volume Part I
Bayesian approach to the pattern recognition problem in nonstationary environment
PReMI'11 Proceedings of the 4th international conference on Pattern recognition and machine intelligence
AICI'11 Proceedings of the Third international conference on Artificial intelligence and computational intelligence - Volume Part I
Improving the performance of data stream classifiers by mining recurring contexts
ADMA'06 Proceedings of the Second international conference on Advanced Data Mining and Applications
A survey on concept drift adaptation
ACM Computing Surveys (CSUR)
A similarity-based approach for data stream classification
Expert Systems with Applications: An International Journal
Concept drift detection via competence models
Artificial Intelligence
| Hi-index | 0.00 |
In their unmodified form, lazy-learning algorithms may havedifficulty learningand tracking time-varying input/output function maps such as those thatoccur in conceptshift. Extensions of these algorithms, such as Time-Windowed forgetting(TWF), can permitlearning of time-varying mappings by deleting older exemplars, but havedecreased classificationaccuracy when the input-space sampling distribution of the learning set istime-varying.Additionally, TWF suffers from lower asymptotic classification accuracy thanequivalentnon-forgetting algorithms when the input sampling distributions arestationary. Other shift-sensitivealgorithms, such as Locally-Weighted forgetting (LWF) avoid the negativeeffectsof time-varying sampling distributions, but still have lower asymptoticclassification innon-varying cases. We introduce Prediction Error Context Switching (PECS)which allowslazy-learning algorithms to have good classification accuracy in conditionshaving a time-varyingfunction mapping and input sampling distributions, while still maintainingtheirasymptotic classification accuracy in static tasks. PECS works by selectingand re-activatingpreviously stored instances based on their most recent consistency record.The classificationaccuracy and active learning set sizes for the above algorithms are comparedin a set oflearning tasks that illustrate the differing time-varying conditionsdescribed above. The resultsshow that the PECS algorithm has the best overall classification accuracyover these differingtime-varying conditions, while still having asymptotic classificationaccuracy competitive withunmodified lazy-learners intended for static environments.