Technology entrepreneurship is closely related to the commercialisation and industrialisation of high technology and is an important engine for a country’s innovation capacity and economic and social development. This study aims to build a systematic theoretical framework and use machine learning algorithms to predict technology entrepreneurship.
Using social cognitive theory, this study explores how individual characteristics, family environment and social environment predict technology entrepreneurship. We utilise data from the Global Entrepreneurship Monitor in 2020 and employ the XGBoost algorithm combined with the Synthetic Minority Over-sampling Technique (SMOTE) algorithm. The performances of several algorithms are compared to determine the optimal predictive model for technology entrepreneurship and test feature importance.
By comparison, it is found that the XGboost algorithm combined with the SMOTE algorithm performs the best in predicting technology entrepreneurship. Furthermore, education, opportunity perception, career choice, social status and media coverage significantly impact technology entrepreneurship.
Technology entrepreneurship involves a combination of several factors. Machine learning algorithms provide a theoretical basis for policymakers to effectively predict and identify technology entrepreneurship. Efforts must be made to increase the education level and improve the perception of opportunities to promote technology entrepreneurship. Additionally, exploring ways to create good social norms is necessary.
This study lays a theoretical foundation for technology entrepreneurship and expands the application of machine learning algorithms in entrepreneurship. By solving complex relationships that are difficult to deal with using traditional statistical models, it can effectively predict and identify technology entrepreneurship and demonstrate the serendipity in technology entrepreneurship.
