| Title | A Combination of Support Vector Machine and Heuristics in On-line Non-Destructive Inspection System |
| Publication Type | Conference Paper |
| Year of Publication | 2018 |
| Authors | Oka, Daisuke, Balage, Don Hiroshan Lakmal, Motegi, Kazuhiro, Kobayashi, Yasuhiro, Shiraishi, Yoichi |
| Conference Name | Proceedings of the 2018 International Conference on Machine Learning and Machine Intelligence |
| Publisher | ACM |
| Conference Location | New York, NY, USA |
| ISBN Number | 978-1-4503-6556-7 |
| Keywords | composability, hammering sound, Metrics, non-destructive inspection, on-line system, pubcrawl, Resiliency, support vector machine, Support vector machines |
| Abstract | This paper deals with an on-line non-destructive inspection system by using hammering sounds based on the combination of support vector machine and a heuristic algorithm. In machine learning algorithms, the perfect performance is hard to attain and it is newly suggested that a heuristic algorithm redeeming this insufficiency is connected to the support vector machine as a post-process. The experimental results show that the combination of support vector machine and the heuristic algorithm attains 100% detection of defective pieces with 18.4% of erroneous determination of non-defective pieces within the upper limit of given processing time. |
| URL | http://doi.acm.org/10.1145/3278312.3278315 |
| DOI | 10.1145/3278312.3278315 |
| Citation Key | oka_combination_2018 |
A Combination of Support Vector Machine and Heuristics in On-line Non-Destructive Inspection System