LAUSR

The study of technological forecasting is an important part of patent analysis. Although fitting models can provide a rough tendency of a technical area, the trend of the detailed content within the area remains hidden. It is also difficult to reveal the trend of specific topics using keyword-based text mining techniques, since it is very hard to track the temporal patterns of a single keyword that generally represents a technological concept. To overcome these limitations, this research proposes a topic-based technological forecasting approach, to uncover the trends of specific topics underlying massive patent claims using topic modelling. A topic annual weight matrix and a sequence of topic-based trend coefficients are generated to quantitatively estimate the developing trends of the discovered topics, and evaluate to what degree various topics have contributed to the patenting activities of the whole area. To demonstrate the effectiveness of the approach, we present a case study using 13,910 utility patents that were published during the years 2000 to 2014, owned by Australian assignees, in the United States Patent and Trademark Office (USPTO). The results indicate that the proposed approach is effective for estimating the temporal patterns and forecast the future trends of the latent topics underlying massive claims. The topic-based knowledge and the corresponding trend analysis provided by the approach can be used to facilitate further technological decisions or opportunity discovery.