Path of Technology Opportunity Identification based on Outlier Patents

Given that the important information of technological innovation contained in outlier patents can provide first-hand data for enterprises to carry out innovation activities efficiently, we rely on the multi-dimensional technological innovation map to construct a path to identify technological opportunities based on outlier patents. [Methods]Firstly, the BERT-LOF algorithm (Local Outlier Factor) is used to extract candidate outlier patents, and the entropy weight method is used to assign weights to the indicators of novelty, scope of application, and development capability and calculate the comprehensive assessment value of candidate outlier patents, so as to filter out the outlier patents with potential value; secondly, the BERTopic theme modeling technique is used to extract technology innovation information from the outlier patents, and the BERTopic theme modeling technique is used to extract technology innovation information from the outlier patents. Secondly, BERTopic theme modeling technology is used to extract technology themes and innovation elements from the outlier patents, and the innovation elements are divided into dimensions based on the multi-dimensional technology innovation map, which is then coupled with multiple innovation laws to identify multiple figurative technological opportunities; lastly, 3D printing technology is taken as an example to illustrate the process of the application of this technological opportunity identification pathway. [Significance] BERT-LOF outlier detection model can effectively improve the extraction accuracy of outlier patents, and its technology opportunities can provide scientific decision-making reference basis for enterprises to improve innovation efficiency and reduce innovation costs.