A Classification Method of the Incident By Extraction From Text
Kazuhiro Morita1, Kenzaburo Yashiro2, Masao Fuketa3
1Kazuhiro Morita, Tokushima University, Japan.
2Kenzaburo Yashiro, Tokushima University, Japan.
3Masao Fuketa, Tokushima University, Japan.
Manuscript received on 06 December 2019 | Revised Manuscript received on 20 December 2019 | Manuscript Published on 31 December 2019 | PP: 1-5 | Volume-8 Issue-12S2 October 2019 | Retrieval Number: L100110812S219/2019©BEIESP | DOI: 10.35940/ijitee.L1001.10812S19
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open-access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: In this research, in order to use for prediction of the accident risk which prevents serious accident and disaster, the method of detecting and classifying an incident from a text is proposed. A multi-attribute matching machine is used for detection and a classification. The feature expression is extracted from the incident case sentence currently released, and detection of an incident and the classification of an accident kind are carried out by the matching rule created from extraction data. Although classification precision was mostly as good as 0.783 as a result of the evaluation experiment, the room for an improvement for extraction precision was seen. The incident which was able to be managed with flawlessness or a slight injury although it was likely to get injured can warn of a big accident, and can urge evasion of it. Therefore, this research which leads to an early warning by detecting and classifying mechanically is meaningful. A future subject is an improvement of extraction precision.
Keywords: Incident Detection, Incident Classification, Multi-Attribute Pattern Matching.
Scope of the Article: Text Mining