Linking Industry, Technology and Business: A Deep Learning-based Approach to Predict their Future Relationships to Identify Innovation Opportunities

Patents and trademarks play distinct yet complementary roles in the realms of industry and innovation. Understanding the intersection of technology and business is vital in today's fast-paced world to identify critical innovative opportunities. However, existing research tends to separately analyze these elements which lacks a holistic perspective. Therefore, this study constructed the linkages between industry, patent, and trademark codes, which is used to analyze their relationship to investigate the innovation path in a holistic perspective. To achieve this goal, first, the deep learning classifier is employed to link patents and trademarks with industries respectively and the link prediction method is employed to link patents and trademarks connected by industry, establishing a refined interconnected network. Next, by using the established network, a multilayer network is constructed with the linkage probabilities of three factors and the link prediction is conducted for each layers, using the artificial node of a target firm. Lastly, the opportunity nodes that can be adjustable to the target firm are filtered in intra and inter layers perspective. This is one of the earliest attempts to take a consolidated view of the interconnected relationships among technology, market, and industry. Also by defining opportunities as the strategic points within the network where a firm can leverage new technological capabilities, market positions, or industry standards to gain competitive advantage, this data-driven linkage analysis enables precise macro-level analysis of innovation opportunities.