مقاله انگلیسی رایگان در مورد شناسایی زودهنگام ثبت اختراعات – الزویر 2019

One of the most challenging problems in technological forecasting is to identify as early as possible those technologies that have the potential to lead to radical changes in our society. In this paper, we use the US patent citation network (1926–2010) to test our ability to early identify a list of expert-selected historically significant patents through citation network analysis. We show that in order to effectively uncover these patents shortly after they are issued, we need to go beyond raw citation counts and take into account both the citation network topology and temporal information. In particular, an age-normalized measure of patent centrality, called rescaled PageRank, allows us to identify the significant patents earlier than citation count and PageRank score. In addition, we find that while high-impact patents tend to rely on other high-impact patents in a similar way as scientific papers, the patents’ citation dynamics is significantly slower than that of papers, which makes the early identification of significant patents more challenging than that of significant papers. In the context of technology management, our rescaled metrics can be useful to early detect recent trends in technical improvement, which is of fundamental interest for companies and investors.

Introduction

While many inventions are granted a patent, only a small fraction of them represent “important” technological advances or will have a significant impact on the market. As a result, a key problem in technological forecasting is to detect which patents are important as early as possible. The literature has designed various indicators of patent importance based on patent data analysis, and it has been found quite consistently (see Section 2) that at least on average, important patents tend to receive more citations. However, this relationship is typically noisy, which suggests that more sophisticated metrics could outperform simple citation count in identifying important patents. Importantly, it takes time for a patent to accumulate citations, which implies that simply counting the number of citations received by a patent may be effective to uncover old important patents, but not to detect important patents shortly after they are granted. In this paper, we propose a network-based metric that identifies important patents better and earlier than citation count. Our metric, time-rescaled PageRank, was introduced by Mariani et al. (2016) to identify expert-selected important papers in physics. It is built on Google’s PageRank algorithm (Brin and Page, 1998) by requiring that node score is not biased by node age. This metric is computationally efficient and thus can be applied on very large datasets (Vaccario et al., 2017). Here we validate this metric on the US patent citation network (1926–2010), by evaluating its ability to detect the expert-selected “important” patents from Strumsky and Lobo (2015). We find that Google’s PageRank outperforms raw citation count in identifying the important patents, which supports the idea that important patents tend to be cited by other important patents.