COA: Finding Novel Patents through Text Analysis

In the last two decades, the value of patents have increased enormously. As a result, companies are showing higher propensity to patent their invention and to strengthen their patent portfolio. However, large portfolios are difficult to manage. One major objective in portfolio management is to rank the patents in terms of their values. Since intellectual property (IP) attorneys’ time is very expensive, an automated or semi-automated software system that expedites and assists the ranking process would be of great value. The existing software systems, targeted at IP professionals, mostly provide web-based services, data feed, advanced patent search interfaces etc. These are very helpful to commend a prior art search or to obtain answers to basic patent related inquiries but are not adequate to assess the value of a patent. Through our research, we build a patent ranking software, named COA (Claim Originality Analysis) that rates a patent based on its novelty. It computes novelty by measuring the impact and the recency of the important phrases that appear in the “claims” section of a patent. In our experiments, we found that COA produces meaningful ranking when comparing it with other indirect patent evaluation metrics— citation count, patent status, and attorney’s rating. In real-life settings, this tool was used by beta-testers in the IBM IP department. Lawyers found it very useful in patent rating, specifically, in highlighting potentially valuable patents in a patent cluster. In this article, we describe the ranking techniques and system architecture of COA. We also present the results that validate its effectiveness.