Search

Patent analysis depicts the innovative directions of research and development (R&D) and technical breakthroughs in given domains, because patents are effective means of intellectual property right (IPR) protection for innovations with viable commercial applications. Patent analysis often helps companies monitor IPs and assess their competitive positions in the marketplace. Patent clustering is one of the crucial analytical approaches in the overall patent analysis procedure for gathering the relevant technologies into coherent and specific clusters. Prediction methods for future technology directions are then utilized to formulate effective R&D strategies and to forecast likely outcomes. Currently available approaches in technology forecasting through clustering are mostly based on general patent text mining to identify key words/terms in the patents and group patents according to similar appearances of keywords. These approaches are not particularly effective at categorizing patent features. A company’s R&D or IP decision makers cannot fully rely on the categorization analysis outcomes to obtain precise insights into emerging technological trends. This paper proposes a new approach to patent claim-based technology clustering to forecast the frontiers of IP protected technologies. The proposed approach is applied for strategic patent analysis of Laser Interference Lithography (LIL) innovations. Patentable features are identified from annotated elements within independent claims of patents. These features are then summarized into simplified sentences for the definition of specific clusters. This method establishes a unique clustering principle to improve patent analysis accuracy and credibility of forecasting future R&D trends based on crucial patent claims.

Analysis of patents reveals trends in technology development in given domains, particularly for commercial adaptation concerning intellectual property (IP) protection. Such an approach allows enterprises to track IPs and evaluate their potential competitiveness vis-à-vis their competitors. Patent clustering is a core step of the entire patent analysis process for conducting technology analysis and can be utilized to group various patents into relatively consistent categories. Forecasting methods are used to develop optimal R&D strategies and anticipate potential outcomes. However, current clustering methods for such technology forecasting, based on general patent keywords or text mining have difficulty in carrying out categorization efficiently and precisely to provide decision makers with insights into technological trends. This research develops a new methodology of patent claim-based technology clustering to predict IP protected technology frontiers as the critical references for the strategic innovation planning. An in-depth patent analysis case study is conducted on two photon polymerization (TPP) technology to demonstrate the generalized methodology working in practice. With annotated elements in independent claims of patent documents, patentable features are identified for given patents. The patentable features to achieve the task of simplification have been highlighted. Afterward, patents are clustered in accordance to the identification of patentable features in the simplified sentences to provide the clusters with names and definitions. The proposed approach establishes a unique clustering principle so as to enhance the accuracy and credibility of patent analysis based on the legally protected patent claims. This approach provides insight into the landscape of future technological trends, particularly for TPP technologies. [ABSTRACT FROM AUTHOR]