Learning Future Classifiers without Additional Data
- 1 January 2018
- journal article
- Published by Japanese Society for Artificial Intelligence in Transactions of the Japanese Society for Artificial Intelligence
- Vol. 33 (2), D-H92_1-H92_1
- https://doi.org/10.1527/tjsai.d-h92
Abstract
We propose probabilistic models for predicting future classifiers given labeled data with timestamps collected until the current time. In some applications, the decision boundary changes over time. For example, in activity recognition using sensor data, the decision boundary can vary since user activity patterns dynamically change. Existing methods require additional labeled and/or unlabeled data to learn a time-evolving decision boundary. However, collecting these data can be expensive or impossible. By incorporating time-series models to capture the dynamics of a decision boundary, the proposed model can predict future classifiers without additional data. We developed two learning algorithms for the proposed model on the basis of variational Bayesian inference. The effectiveness of the proposed method is demonstrated with experiments using synthetic and real-world data sets.Keywords
This publication has 22 references indexed in Scilit:
- A survey on concept drift adaptationACM Computing Surveys, 2014
- COMPOSE: A Semisupervised Learning Framework for Initially Labeled Nonstationary Streaming DataIEEE Transactions on Neural Networks and Learning Systems, 2013
- Reacting to Different Types of Concept Drift: The Accuracy Updated Ensemble AlgorithmIEEE Transactions on Neural Networks and Learning Systems, 2013
- StreamAR: Incremental and Active Learning with Evolving Sensory Data for Activity RecognitionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2012
- Covariate Shift by Kernel Mean MatchingPublished by MIT Press ,2008
- Applying lazy learning algorithms to tackle concept drift in spam filteringExpert Systems with Applications, 2007
- Learning drifting concepts: Example selection vs. example weightingIntelligent Data Analysis, 2004
- Models and issues in data stream systemsPublished by Association for Computing Machinery (ACM) ,2002
- Bayesian parameter estimation via variational methodsStatistics and Computing, 2000
- User Modeling for Adaptive News AccessUser Modelling and User-Adapted Interaction, 2000