해석가능 인공지능을 활용한 바이오화학 기술의 비즈니스 잠재성 평가 (Evaluating the Business Potential of Bio-based Chemical Technologies Using Explainable AI)

서지정보

· 발행기관 : 대한산업공학회
· 수록지 정보 : 대한산업공학회지 / 49권 / 3호 / 223 ~ 236페이지
· 저자명 : 이지호, 이승현, 손은수, 윤장혁, 이재민

초록

Since the fossil fuel-based industry significantly contributes to air pollution and climate change, better living through fossil fuel has come at a cost. In this connection, Bio-based chemical technologies based on reusable biomass such as cells or other living things are receiving great attraction. But at the same time, they are considered as high-risk investments that require a long-term effort to be adopted by businesses. Therefore, building on a common academic consensus that there is a strong correlation between patent lifetime and business potential, this study proposes a machine learning model to predict the lifetime of bio-based chemical technologies. To this end, CAS (Chemical Abstract Service) patent database and PATSTAT (Worldwide Patent Statistical Database) are used to identify global bio-based chemical technology patents. The proposed model identifies bio-based chemical technologies that have high business potential with an accuracy of 81%. Further, the application of an explainable AI algorithm to the model found that the geographical scope of technologies and the size of stakeholders of a business significantly influence the business potential of bio-based chemical technologies. Our research results can be used for the investment and management process for bio-based chemical technologies with high business potential.

영어초록

Since the fossil fuel-based industry significantly contributes to air pollution and climate change, better living through fossil fuel has come at a cost. In this connection, Bio-based chemical technologies based on reusable biomass such as cells or other living things are receiving great attraction. But at the same time, they are considered as high-risk investments that require a long-term effort to be adopted by businesses. Therefore, building on a common academic consensus that there is a strong correlation between patent lifetime and business potential, this study proposes a machine learning model to predict the lifetime of bio-based chemical technologies. To this end, CAS (Chemical Abstract Service) patent database and PATSTAT (Worldwide Patent Statistical Database) are used to identify global bio-based chemical technology patents. The proposed model identifies bio-based chemical technologies that have high business potential with an accuracy of 81%. Further, the application of an explainable AI algorithm to the model found that the geographical scope of technologies and the size of stakeholders of a business significantly influence the business potential of bio-based chemical technologies. Our research results can be used for the investment and management process for bio-based chemical technologies with high business potential.

참고자료

· 없음