Confidential — Stefan Michaelcheck Only

AI-enhanced multi-dimensional measurement of technological convergence through heterogeneous graph and semantic learning

2026model innovationnovelmethod

Siming Deng, Yi Zhang, Mengjia Wu, Chunjuan Luan, Runsong Jia

Scientometrics

https://doi.org/10.1007/s11192-025-05512-xOpenAlex: W7124600053
1
URLs Found
0
Internal Citations
5
Authors
usable
Abstract Quality
GPT-5.5 Abstract Analysis

Problems Identified (2)

Technological convergence measurement challenge: Accurately measuring technological convergence is difficult because convergence is multidimensional and evolves over time.

Technological convergence measurement challenge: Accurately measuring technological convergence is difficult because convergence is multidimensional and evolves over time.

Proposed Solutions (5)

AI-enhanced Technological Convergence Index: The study develops an AI-enhanced Technological Convergence Index that measures technological convergence through depth and breadth dimensions.

Heterogeneous graph and semantic depth modeling: The depth dimension uses IPC textual descriptions and patent metadata in a heterogeneous graph modeled with HGT and SBERT to represent knowledge integration across technological boundaries.

Shannon diversity breadth measurement: The breadth dimension quantifies the diversity of technological fields in patents using the Shannon Diversity Index.

Entropy-weighted index aggregation: The final TCI combines depth and breadth using the Entropy Weight Method to assign objective weights based on information entropy.

AI-enhanced Technological Convergence Index: The study develops an AI-enhanced Technological Convergence Index that measures technological convergence through depth and breadth dimensions.

Results (3)

Comparative advantages over established measures:

Empirical reliability via robustness test:

Positive effect on patent quality:

Research Domain

Technological convergence measurement; patent analytics; scientometrics

← Back to all papers