A stochastic patent citation analysis approach to assessing future technological impacts

Abstract

The anticipation and forecast of technological changes are of vital importance, as technological advances become increasingly fast and complex. What is at the core is identification of the current technologies that will drive technological changes over the coming few years. In this respect, numerous approaches have been devised to assess future technological impacts based on patent citation information, but do not provide a fair reflection of dynamic and idiosyncratic aspects of technological impacts as they are deterministic methods based on simple citation counts. We propose a stochastic patent citation analysis that can assess future technological impacts in a time period of interest by employing the future citation count as a proxy. At the heart of the proposed approach is a Pareto/NBD (Negative Binomial Distribution) model for taking into account the dynamic and idiosyncratic aspects of technological impacts. A patent citation matrix is first constructed for each time unit with citation patterns of the past. The future technological impacts are then derived by Pareto/NBD sub-model and gamma-gamma sub-model. A case study of the display technology patents is presented to illustrate the proposed approach. We believe our method can be employed in various research fields, from narrow patent valuation, to broad technological analysis and planning.