Adaptive Ensemble Learning With Confidence Bounds

Abstract

Abstract-Extracting actionable intelligence from distributed, heterogeneous, correlated and high-dimensional data sources requires run-time processing and learning both locally and globally. In the last decade, a large number of meta-learning techniques have been proposed in which local learners make online predictions based on their locally-collected data instances, and feed these predictions to an ensemble learner, which fuses them and issues a global prediction. However, most of these works do not provide performance guarantees or, when they do, these guarantees are asymptotic. None of these existing works provide confidence estimates about the issued predictions or rate of learning guarantees for the ensemble learner. In this paper, we provide a systematic ensemble learning method called Hedged Bandits, which comes with both long run (asymptotic) and short run (rate of learning) performance guarantees. Moreover, our approach yields performance guarantees with respect to the optimal local prediction strategy, and is also able to adapt its predictions in a data-driven manner. We illustrate the performance of Hedged Bandits in the context of medical informatics and show that it outperforms numerous online and offline ensemble learning methods.

Budget-Constrained Edge Service Provisioning With Demand Estimation via Bandit Learning

“…Though the algorithm parameter h T and the regret upper bound is given based on a known time horizon T , COERR can be easily extended to work with unknown time horizon with the assistance of doubling-trick [36], [37]. The key idea of doubling-trick is to partition the time into multiple phases (j = 1, 2, 3, ...) with doubling length (T 1 , T 2 , · · · ), e.g., if the length of phase is T 1 = T , then the length of j-th phase is 2 j−1 T .…”

Expert advice ensemble for thyroid disease diagnosis

“…The real-time classification algorithms have emerged that analyze the decision making process and make improvements in real-time. To provide an expert system, one needs to answer the following questions online [14]: What will be the cost for wrong decision? What will be the weights for different classifiers?…”

Actionable intelligence and online learning for semantic computing

“…These streams are then mined in realtime to assist numerous semantic computing applications (see Fig. 1): patient monitoring, 1 personalized diagnosis, 2,3 personalized treatment recommendation, 4,5 recommender systems, [6][7][8] social networks, 9 network security, 10 multimedia content aggregation, 11 personalized education 12,13 etc. Hence, online data mining systems have emerged that enable such applications to analyze, extract actionable intelligence and make decisions in real-time, based on the correlated, highdimensional and dynamic data captured by multiple heterogeneous data sources.…”