Scholarlink

<p>Gradient boosting has been extensively studied in batch learning. Recently, its streaming adaptation, Streaming Gradient Boosted Trees (<span>Sgbt</span>), has surpassed existing state-of-the-art random subspace and random patches methods for streaming classification under various drift scenarios. However, its application in streaming regression remains unexplored. Vanilla <span>Sgbt</span> with squared loss exhibits high variance when applied to streaming regression problems. To address this, we utilize bagging streaming regressors in this work to create Streaming Gradient Boosted Regression (<span>Sgbr</span>). Bagging streaming regressors are employed in two ways: first, as base learners within the existing <span>Sgbt</span> framework, and second, as an ensemble method that aggregates multiple <span>Sgbt</span>s. Our extensive experiments on 11 streaming regression datasets, encompassing multiple drift scenarios, demonstrate that the <span>Sgb(Oza)</span>, a variant of the first <span>Sgbr</span> category, significantly outperforms current state-of-the-art streaming regression methods in terms of both predictive power and computational cost.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Gradient boosted bagging for evolving data stream regression

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年8月6日

查看全文
  • Nuwan Gunasekara,
  • Bernhard Pfahringer,
  • Heitor Murilo Gomes,
  • Albert Bifet

摘要

Gradient boosting has been extensively studied in batch learning. Recently, its streaming adaptation, Streaming Gradient Boosted Trees (Sgbt), has surpassed existing state-of-the-art random subspace and random patches methods for streaming classification under various drift scenarios. However, its application in streaming regression remains unexplored. Vanilla Sgbt with squared loss exhibits high variance when applied to streaming regression problems. To address this, we utilize bagging streaming regressors in this work to create Streaming Gradient Boosted Regression (Sgbr). Bagging streaming regressors are employed in two ways: first, as base learners within the existing Sgbt framework, and second, as an ensemble method that aggregates multiple Sgbts. Our extensive experiments on 11 streaming regression datasets, encompassing multiple drift scenarios, demonstrate that the Sgb(Oza), a variant of the first Sgbr category, significantly outperforms current state-of-the-art streaming regression methods in terms of both predictive power and computational cost.