| Full text | |
| Author(s): |
Cassales, Guilherme
[1]
;
Gomes, Heitor
[2]
;
Bifet, Albert
[2]
;
Pfahringer, Bernhard
[2]
;
Senger, Hermes
[1]
Total Authors: 5
|
| Affiliation: | [1] Univ Fed Sao Carlos, Sao Carlos - Brazil
[2] Univ Waikato, Hamilton - New Zealand
Total Affiliations: 2
|
| Document type: | Journal article |
| Source: | INFORMATION SCIENCES; v. 580, p. 260-282, NOV 2021. |
| Web of Science Citations: | 0 |
| Abstract | |
Often, machine learning applications have to cope with dynamic environments where data are collected in the form of continuous data streams with potentially infinite length and transient behavior. Compared to traditional (batch) data mining, stream processing algorithms have additional requirements regarding computational resources and adaptability to data evolution. They must process instances incrementally because the data's continuous flow prohibits storing data for multiple passes. Ensemble learning achieved remarkable predictive performance in this scenario. Implemented as a set of (several) individual classifiers, ensembles are naturally amendable for task parallelism. However, the incremental learning and dynamic data structures used to capture the concept drift increase the cache misses and hinder the benefit of parallelism. This paper proposes a mini-batching strategy that can improve memory access locality and performance of several ensemble algorithms for stream mining in multi-core environments. With the aid of a formal framework, we demonstrate that mini-batching can significantly decrease the reuse distance (and the number of cache misses). Experiments on six different state-of-the-art ensemble algorithms applying four benchmark datasets with varied characteristics show speedups of up to 5X on 8-core processors. These benefits come at the expense of a small reduction in predictive performance. (c) 2021 Elsevier Inc. All rights reserved. (AU) | |
| FAPESP's process: | 18/22979-2 - IoT-SED: security and efficiency in data transport on Internet of Things |
| Grantee: | Daniel Macêdo Batista |
| Support Opportunities: | Regular Research Grants |
| FAPESP's process: | 19/26702-8 - Trends on high performance computing, from resource management to new computer architectures |
| Grantee: | Alfredo Goldman vel Lejbman |
| Support Opportunities: | Research Projects - Thematic Grants |
| FAPESP's process: | 15/24461-2 - A study of business models for the federation of services supporting e-Science |
| Grantee: | Francisco Vilar Brasileiro |
| Support Opportunities: | Research Projects - Thematic Grants |