A transferability evaluation model for intellectual property

Abstract

As a type of intellectual property rights, patents are a vast source of human-generated technological knowledge; as such, patent evaluations from various perspectives have long been of primary interest to researchers. However, although patents are transferable assets with economic and technological value, limited attention has been paid to the possibility of realizing a patent’s potential through its transaction. Consequently, this study develops a patent transferability evaluation model by applying deep neural networks (DNNs) to various patent indicators and the corresponding historical patent rights transaction data. To this end, this study (1) constructs a patent database with patents and their corresponding historical patent rights transaction data from the Korean Patent Office database; (2) defines how to extract a variety of patent indicators related to patent transferability that do not depend on forward citations; (3) builds a patent transferability evaluation model based on DNN; and (4) validates the performance and effectiveness of the developed model. This study contributes to the literature by being one of the first to quantitatively evaluate patents in terms of transferability and, thus, the proposed model can be used for valuing patents and distinguishing quality patents that are marketable.