Forecasting Technology Emergence from Metadata and Language of Scientific Publications and Patents

This paper describes a multidisciplinary study and development effort to analyze full text and metadata of scientific articles and patents for indicators of new disruptive and game-changing technical breakthroughs. The system we are developing can scan millions of documents in two languages, English and Chinese, and extract meaningful trends and predictions. Whereas traditional approaches to innovation analytics rely on citation analysis to analyze impact or identify the most influential patents or researchers in the field, our system takes a step further and combines these methods with an analysis of text in order to identify and characterize emerging technologies. The paper describes the indicators and forecasting models, as well as presents the results of applying these indicators to forecast levels of interest in a particular technology based on the analysis of English and Chinese patents. It further shows how the indicators we developed can provide insights into the nature and the lifecycle of emerging technologies. Conference Topic Indicators