A comparative study of the leading machine learning techniques and two new optimization algorithms |
| |
Authors: | P Baumann DS Hochbaum YT Yang |
| |
Institution: | 1. Department of Business Administration, University of Bern, Schützenmattstrasse 14, Bern 3012, Switzerland;2. Industrial Engineering and Operations Research Department, University of California, Berkeley, CA 94720, USA;3. Amazon.com, 101 Main Street, Cambridge, MA 02142, USA |
| |
Abstract: | We present here a computational study comparing the performance of leading machine learning techniques to that of recently developed graph-based combinatorial optimization algorithms (SNC and KSNC). The surprising result of this study is that SNC and KSNC consistently show the best or close to best performance in terms of their F1-scores, accuracy, and recall. Furthermore, the performance of SNC and KSNC is considerably more robust than that of the other algorithms; the others may perform well on average but tend to vary greatly across data sets. This demonstrates that combinatorial optimization techniques can be competitive as compared to state-of-the-art machine learning techniques. The code developed for SNC and KSNC is publicly available. |
| |
Keywords: | Data mining Supervised machine learning Binary classification Comparative study Supervised normalized cut |
本文献已被 ScienceDirect 等数据库收录! |
|