Study on the Prediction of Patent Hiding Company Using Patent Information Analysis

Establishment of an effective R&D strategy can be achieved through Prior art search. Prior art search is aimed at identifying major applicants in a specific technology field and is mainly carried out using patents. However, some companies hide their patents as M&A or other companies with most shares in order to avoid exposure. It is difficult to identify the actual owners of patents of such patent hiding companies. Thus, patent hiding companies hinder successful prior art search. To solve this problem, this paper proposes a model for predicting patent hiding companies by using text information and quantitative indicators of patents. As experimental data, N screen technology patents registered in USPTO are used. As a result of the experiment, Cleversafe was found to be IBM's patent hiding company.