특징 추출 및 그래프 기반 준지도학습을 활용한 중소기업 성장변화 분석: 코로나19 전후의 반도체 산업을 중심으로 = SME Growth Analysis Using Feature Extraction and Graph-Based Semi-Supervised Learning: A Case Study of the Semiconductor Industry Before and After COVID-19

다국어 초록 (Multilingual Abstract)

The pandemic of COVID-19 has caused a serious economic shock worldwide, and the domestic SMEs community also suffered from various aspects, including financing and securing demand sources. Accordingly, it is necessary to identify potential growth opportunities and threats of SMEs. This study develops an AI-based model to diagnose changes in the growth potential of the domestic semiconductor industry before and after COVID-19. Various information was collected and utilized, and the growth potential of each company is predicted by combining feature extraction and graph algorithm-based AI techniques. As a result of the research, the resilience and growth potential of the semiconductor industry after COVID-19 showed a big difference depending on the type of product and the region to which it belongs. Companies in the semiconductor application products sector showed stable growth rather than semiconductor devices or test equipment. Moreover, regions with larger semiconductor industry ecosystem were more favorable for company's recovery and growth. In theory, this study is differentiated in that it can objectively diagnose complex characteristics of the semiconductor industry by learning nonlinear relationships between various variables, unlike previous studies that mainly dealt with traditional indicators and statistical analysis. In addition, since the research results empirically analyzed the impact of COVID-19 on products, industries, regions, and corporate types, it is expected that the government or companies will contribute to establishing practical alternatives and strategies.