MARLINE: Multi-Source Mapping Transfer Learning for Non-Stationary Environments
MARLINE: Multi-Source Mapping Transfer Learning for Non-Stationary Environments
复制标题
DOI：
10.1109/icdm50108.2020.00021
复制
发表时间：
2020-11
期刊：
影响因子：
--
通讯作者：
Honghui Du;Leandro L. Minku;Huiyu Zhou
中科院分区：
文献类型：
--
作者：
Honghui Du;Leandro L. Minku;Huiyu Zhou
概念漂移是在线学习中的一个主要问题，因为它影响数据流挖掘系统的预测性能。最近的研究已经开始探索来自不同来源的数据流，作为解决给定目标领域中概念漂移的一种策略。这些方法假设至少一个源模型表示一个类似于目标概念的概念，这在许多真实世界的场景中可能不成立。在本文中，我们提出了一种新的方法，称为非平稳环境下的带转移学习的多源映射(Marline)。即使源和目标概念不匹配，Marline也可以从非静态环境中的多个数据源的知识中受益。这是通过将目标概念投影到每个源概念的空间来实现的，使得多个源子分类器能够作为集合的一部分对目标概念的预测做出贡献。在几个合成和真实世界的数据集上的实验表明，Marline比几种最先进的数据流学习方法更准确。
Concept drift is a major problem in online learning due to its impact on the predictive performance of data stream mining systems. Recent studies have started exploring data streams from different sources as a strategy to tackle concept drift in a given target domain. These approaches make the assumption that at least one of the source models represents a concept similar to the target concept, which may not hold in many real-world scenarios. In this paper, we propose a novel approach called Multi-source mApping with tRansfer LearnIng for Nonstationary Environments (MARLINE). MARLINE can benefit from knowledge from multiple data sources in non-stationary environments even when source and target concepts do not match. This is achieved by projecting the target concept to the space of each source concept, enabling multiple source sub-classifiers to contribute towards the prediction of the target concept as part of an ensemble. Experiments on several synthetic and real-world datasets show that MARLINE was more accurate than several state-of-the-art data stream learning approaches.