Knowledge source strategy and enterprise innovation performance: dynamic analysis based on machine learning

ABSTRACT

By employing ergodic theory and applying the most advanced machine-leaning methods, this study exploits the rules of multi-dimensional, phased and non-linear dynamic evolution between the breadth and depth of knowledge sources and the innovation performance. The following conclusions are obtained. First, regarding explorative innovation, when both the breadth and depth of the knowledge source are at a low level, the enhancement of the breadth of the knowledge source may rapidly lift explorative innovation performance; when the knowledge source is at a high level, the theory of ‘ambidexterity balance’ is more applicable to find a balance between the breadth and the depth of the knowledge source for the enhancement of explorative innovation performance. Second, in terms of exploitative innovation, ‘ambidexterity balance’ theory can be applied at all levels. In other words, the balance of the breadth and the depth of the knowledge sources greatly enhances the exploitative innovation performance.

KEYWORDS:

• Knowledge source strategy
• innovation performance
• machine learning
• dynamic evolution

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Xin Jin is an Assistant Professor at School of Economics & Management, Shanghai Maritime University in China.

Jie Wang is a Consulting Professor at Center for Sustainable Development and Global Competitiveness, Stanford University.

Tianshu Chu is currently the PhD student at the Civil and Environmental Engineering department of Stanford University.

Jinghua Xia is President of the Research Institute of Landray Software Co., Ltd. in China.