Developing a predictive model for anticipating technology convergence: A transformer-based model and supervised...

Area of Science:

• Biomedical Engineering
• Health Informatics
• Technology Management

Background:

• Technology convergence is rapidly transforming the bio-healthcare sector.
• Anticipating these convergence trends is crucial for innovation and competitive advantage.
• Existing methods often lack the sophistication to capture complex interrelationships.

Purpose of the Study:

• To develop a robust methodology for predicting technology convergence in bio-healthcare.
• To leverage text mining and supervised learning for identifying emerging trends.
• To provide actionable insights for interdisciplinary collaboration and business strategy.

Main Methods:

• Utilized transformer models (e.g., BERTopic with KeyBERT, OpenAI) for technology topic extraction from patents and research articles.
• Integrated technological similarity, link prediction, and causal relationships as input features for prediction models.
• Employed machine learning and deep learning techniques to train predictive models for technology convergence.

Main Results:

• Successfully identified potential technology convergence opportunities within the bio-healthcare domain.
• Revealed emerging trends and intricate relationships between technology topics.
• Demonstrated the effectiveness of the proposed integrated approach in predicting convergence.

Conclusions:

• The study offers a novel methodology for anticipating future technology convergence in bio-healthcare.
• Findings provide a foundation for identifying interdisciplinary collaboration opportunities.
• The approach has significant implications for businesses aiming for sustainable growth through innovation.