The spring up of new & emerging technologies brings a lot of innovation opportunities for society, which enables technology opportunities analysis attracts increasing attention by both industry and academia recently. This study proposes a hybrid approach which integrates topic modeling, sentiment analysis, patent mining and expert judgment to identify technological topics and the potential development opportunities. In order to illustrate how the approach is validated and optimized, and to present its potential to contribute technical intelligence for research and development management, we apply the hybrid approach to analyze a set of 9883 DII records that involved dye sensitized solar cell research. The main contributions of this study include three-fold. First, we distinguished the terms in the different parts of DII patent documents when utilizing them to recognize technical topics. Second, we utilized the terms extracted from the Advantage and the Use part to identify topics on technical problems and applications, and proposed a probability-based topic relation measurement method to identify the relationships of the technical problems and applications with the core sub-technologies. Third, we introduced both topic modeling and sentiment analysis to support technical topic analysis.
For FULL-TEXT of manuscript https://eeke-workshop.github.io/2022/submissions/EEKE2022_paper_8.pdf
Author(s): Tingting Ma, Ruiping Cheng, Hongshu Chen, XiaoZhou
Organization(s): Beijing Wuzi University, Beijing Institute of Technology, Xidian University
Source: 3rd Workshop on Extraction and Evaluation of Knowledge Entities from Scientific Documents (EEKE2022)