Strategic forecasting of internet of things technologies through patent social network and innovation cluster analysis

Abstract

The rapid proliferation of Internet of Things (IoT) technologies necessitates robust forecasting mechanisms to guide strategic decision-making amid increasingly complex innovation landscapes. Despite extensive research employing patent analysis for technology forecasting, existing studies lack systematic integration of social network analysis, advanced text mining, and life cycle modeling to comprehensively map IoT technological evolution and collaborative dynamics. This study addresses these gaps by analyzing 154,227 IoT-related patents through a unified methodological framework combining BERT-based text embeddings, k-means clustering with Davies-Bouldin optimization, S-curve life cycle modeling, and Louvain community detection. The analysis identified nine distinct technology clusters spanning foundational infrastructure (Smart Monitoring and Sensor Systems, Network Communication and Data Transmission) to domain-specific applications (Agricultural IoT, Connected Vehicle Technologies). Life cycle assessment revealed temporal convergence, with eight clusters reaching saturation between 2023 and 2027, reflecting ecosystem-wide synchronization driven by standardization imperatives, platform consolidation, and pandemic-accelerated digital transformation. Social network analysis uncovered five major collaborative communities exhibiting divergent strategic orientations: from extreme specialization (Global Telecommunications Technology Leaders: 87.7% network communication focus) to diversified portfolios (China State Grid IoT Consortium: balanced infrastructure investment). Cross-analysis revealed complementary innovation strategies where infrastructure operators pursue breadth, telecommunications specialists maintain focused expertise, and academic researchers emphasize development-aligned agendas.