Classifying technology patents to identify trends: Applying a fuzzy-based clustering approach in the Turkish textile industry

Economic and technological advances have dramatically changed trends in capital investment. As a result, conventional businesses are at risk of being replaced with businesses that embrace new technology and trends. It is essential to identify and act on business opportunities as they appear within these new trends. This paper introduces an approach that uses patent count data to discover technology trends through a fuzzy-based clustering methodology. Technologies are classified into three types—trendy, classic, and dated—that are identified using K-means clustering. This methodology has been tested using textile patents retrieved from the on-line database of the Turkish Patent Institute. As a result, the most promising sub-sectors in the Turkish textile industry can be determined.