텍스트마이닝을 활용한 전투협조회의의 회의록 분석 방법론 연구
A Study on Methodology for Analyzing Minutes of the Combat Collaboration Meetings Using Text Mining
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
군에서 전투지휘와 관련해 데이터들은 훈련 상황에서 많이 발생되고 있으나, 전투지휘 데이터를 정형화 하거나 그에 대한 분석은 미미한 현실이다. 더욱이 보안상 이유로 전투지휘 데이터를 축적하지도 못하고 있다. 이러한 현실 속에서도 AI 참모 또는 참모지원 AI에 대한 연구의 필요성은 국방혁신 4.0 추진과 함께 많이 요구되고 있다. 그러나 이러한 요구사항을 구현하기 위해서는 먼저 전투지휘 데이터를 정형화하는 것이 필요하다. 본 연구에서는 지휘관 및 참모의 의사결정과정 상에 회의를 통해서 생산되는 전투지휘 데이터를 정형화시키고, 정형화된 전투지휘정보를 분석하는 방법론을 제시하였다. 보안상의 이유로 전투지휘 데이터를 실제 획득할 수가 없어서 이와 유사한 형태의 국방 관련 텍스트 데이터인 국회 국방위원회 회의록으로 연구내용을 적용하였다. 국회 국방위원회 회의록 텍스트 데이터에 텍스트 전처리와 핵심단어구축, 토픽모델링, 네트워크분석, 감성분석, t-SNE 군집화 등의 텍스트 마이닝을 적용하여 텍스트 데이터를 정형화하고 분석하는 방법론을 제안하였다. 본 연구에서 제안하는 방법론으로 회의의 주제들을 도출하고, 화자별 발화량 비교와 감성분석을 통해서 회의를 주재하는 주관자(군 회의에서는 지휘관)의 회의 진행 스타일에 대해서 평가할 수 있는 방안을 제시하였다. 이를 통해서 보다 효과적이고 생산적인 전투지휘가 어떤 것인지를 분석하여 조언해 줄 수 있는 AI 참모 또는 참모지원 AI 개발의 여건을 만드는데 기여할 수 있을 것이다.
Abstract
In the military field, data related to combat command is often generated during training situations, yet there has been minimal effort in structuring and analyzing actual battle command information. Additionally, the accumulation of battle command information is greatly limited due to security reasons. Despite these constraints, there is a growing demand for research on AI systems supporting staff officers in decision-making, particularly with the continuous advancement of Defense Innovation 4.0. To address this, there is a need for research to first structure battle command information to extract meaningful insights. This study presents a methodology, along with case studies, for structuring and deriving meaningful information from battle command data generated during combat coordination meetings involving commanders and staff, aiming to provide insights necessary for decision support systems. As acquiring actual combat coordination meeting content was restricted due to military security concerns, case analysis was conducted using minutes from National Defense Committee meetings, which exhibit similar patterns to combat coordination meetings. The proposed methodologies for structuring battle command information enable comparative analysis with operational orders to understand how battle command information impacts actual combat outcomes. Moreover, analysis techniques such as sentiment analysis allow assessment of the efficiency with which commanders lead discussions over time, understanding the inclinations of meeting participants, and identifying decision-making patterns through topic modeling. Clustering of participants' speech content allows for the identification of similar speakers, facilitating comparison of thoughts among individuals. This study not only presents methodologies for structuring and analyzing battle command information crucial for developing AI decision support systems but also suggests directions for constructing such systems.
참고문헌 (29)
- Army Headquarters, "Field Regulations 1-1 Military Terminology Dictionary" (2019)
- Arun, R., "On finding the natural number of topics with latent dirichlet allocation: Some observations", Advances in Knowledge Discovery and Data Mining: 14th P acific-Asia Conference, 14 (2010), Springer Berlin Heidelberg
- Bang, H., "A Methodology for Temporally Representing the Major Topics of Texts Using Text Mining", J ournal of the Korean Data I nformation Science Society, 30(6), pp.1259~1276 (2019)
- Blei, D. M., "Latent dirichlet allocation", J ournal of machine Learning research, 3, pp.993~1022 (2003)
- Choi, K. W., "Hangeul Stem Extraction Algorithm for Text Mining Based on Natural Language Processing", Proceedings of the Korean Institute of Information and Commucation Sciences Conference, pp.718~721 (2017), The Korea I nstitute of I nformation and Commucation Engineering
- "국방 텍스트마이닝" (2024)[URL]
- Fotis B., "Collection and Monitoring via Planning for Active Situational Scenarios(COMPASS)" (2020), NNSA
- Griffiths, T. L., "Finding scientific topics", National Acad Sciences, pp.5228~5235 (2004)
- Han, C., "Data Modeling Approach for Decision Conditions in Building AI Staff", Journal of Internet Computing and Services, 21(1) (2020)
- Han, C., "Construction Approach of Defense AI Command and Control Platform", Journal of the Korean Institute of Communication Sciences, 44(4) (2019)
- Joint all-domain command and control:Background and issues for Congress, "Hoehn, J. R.", Congressional Research Service, ~, pp.Congressional Research Service , 46725:18 (2021)
- Hong, Se Eun, "Exploring the similarity between incident types based on text mining and cluster analysis through vectorizing contents of emergency calls", Korea National P olice University, 20(3), pp.63~86 (2020)
- Jung, J., "A Study on User-customized Housing Design Platform as a Decision Support System", Journal of Convergence Technology, 3(1), pp.15~20 (2019)
- Jo, S., "Analysis of Trends and Implications in Defense Policy Research Using Topic Modeling and Factor Analysi", Military Operations Research Society of Korea, 49(3), pp.61~74 (2023)
- Kam, M., "A Study on Differences of Contents and Tones of Arguments among Newspapers Using Text Mining Analysis", J I ntell I nform Syst, 18(3), pp.53~77 (2012)
- Kang D., "study on media reports on self-employment in the era of COVID-19 using text mining techniques", Master, pp.2022
- Kim, K., "A review on the t-distributed stochastic neighborsembedding", The Korean Journal of Applied Statistics, 36(2), pp.167~117 (2023)
- Kwak, S., "Keywords and Topic Analysis of Social Issues on Twitter Based on Text Mining and Topic Modeling", TKIP S Trans. Softw. and Data Eng, 8(1), pp.13~18 (2019)
- A Study on the Establishment of a Big-data Platform to Support Policy Decision-making in the Port Sector, "Lee, K.", Korea Maritime I nstitute, ~, pp.Korea Maritime I nstitute . 2021-13 (2021)
- Lee, D., "Development of Command Decision Support System Based on Artificial Intelligence (AI) and M&S", Defense Forum (2017)
- Lee, M., "Study on Automatic Prediction Methodology for AI Information Staff Penetration by Fusion of Terrain Information and Cognitive Information", Korean Journal of Military Science, 79(1) (2023)
- Lee C., "Study on analysis of North Korea`s news trends associated with provocations using text mining", Research I nstitute On National Security Affairs, 59(4), pp.103~124 (2016)
- Lee, C., "Trends in AI Military Staff Technology for Decision Support of Commanders", Trends in Electronics & Telecommunication Analysis, 36(1) (2021)
- Le, Q., "Distributed representations of sentences and documents", I nternational conference on machine learning. PMLR, pp.1188~1196 (2014)
- Ministry of National Defense, "Defense Innovation 4.0", ~ (2023)
- Automatic Labeling of Equipment Data Using t-SNE, "Moon, S.", Korean I nstitute of I ndustrial Engineers, 576~580 (2023)
- Moon, H., "Research on Development Directions of Army Scientific Combat Training (KCTC) Based on Big Data Analysis of Training Results: Objective Diagnosis and Evaluation of Training Units", Defense Science and Technology, 7(4) (2021)
- Park, Y., "Deep Learning Using TensorFlow Keras" (2023), Free Academy