Abstract— Because the domain of nuclear power is highly specialized and complex, human experts have been utilized to manually evaluate all the documents submitted for export permission, causing the evaluation process to be slow and costly. Toward alleviating the problem of relying on laborious and costly human experts, the present research examines alternative approaches of text categorization, which is a key component of the case-based reasoning system proposed for the retrieval of documents only in the classes where a new export request case is related. Specifically, we examined three text categorization approaches: 1) manual approach involving a field expert, 2) automatic approach utilizing the TF-IDF scheme, and 3) semi-automatic approach involving both student experts and the TF-IDF scheme. Among the three methods, semi-automatic approach is the most efficient and effective in extracting keywords, demonstrating that the combination of machine and human is a promising solution that can effectively overcome the issues of expertise scarcity, time, cost, and accuracy simultaneously.
Index Terms— Nuclear exports control system, case-based reasoning, text categorization, keyword extraction.
Uihyun Kim is with Tibero, Republic of Korea.
Hyunji Kim and Mun Yi are with the Department of Knowledge Service Engineering at KAIST, Republic of Korea (e-mail: munyi@kaist.ac.kr).
Donghoon Shin is with Korea Institute of Nuclear Nonproliferation and Control (KINAC), Republic of Korea.
[PDF]
Cite: Uihyun Kim, Hyunji Kim, Mun Yi, and Donghoon Shin, " Nuclear Exports Control System Using Semi-Automatic Keyword Extraction," International Journal of Information and Electronics Engineering vol. 4, no. 4, pp. 293-297, 2014.
