How does AI Select the Best Model of Business Innovation Prediction?

To overcome the global economic recession, many companies are seeking to strengthen their innovation capabilities. Although the academic world focuses on identifying factors that determine the innovation capabilities using traditional quantitative methodologies, the need to prepare strategies that reflect high-level structures closely related to real-world problems is emerging. This study identified the AI algorithm for business innovation performance prediction, and proposed strategies required at the present time, focusing on important factors. This study utilized data from the 2022 Korean Innovation Survey (KIS) in the Service Sector, and the performance of optimal tree ensemble-based machine learning was compared. As a result of the analysis, the prediction performance of Soft Voting with weighted XGBoost was the best than single algorithms. Also, important factors related to the company's internal funding, internal R&D, and customer-tailored focus strategies were derived. This study has practical implications in that it proposes corporate funding deregulation and internal strategies as well as academic implications for selecting an algorithm that understands and predicts the overall mechanism of innovation performance. Also, this approach would serve as basic data to prepare innovation strategies for economic growth.