Identifying Technological Trends: A Patent Analysis Method for Technology Forecasting

Patents are extensive and reliable sources of data on technological inventions, serving as the basis for patent analysis tasks. Among these tasks, technology forecasting is essential for research, development and decision-making in organizations. This paper proposes a decision-making support method capable of identifying technological trends. To achieve this, we explore the learning of network representations by applying link prediction algorithms to identify potential trends in the links between technologies. To demonstrate the effectiveness of the proposed method, we conducted experiments in the field of carbon technology. Our link prediction model reached a mean performance of 0.91, considering the ROC-AUC metric.