Fig 1.
The proposed research framework for identifying technology convergence.
Table 1.
Overview of data gathering details.
Fig 2.
Applying the BERTopic process for generating technology topic.
Fig 3.
Exploration of technological similarity between technology topics.
Fig 4.
Constructing technology topic networks and applying link prediction measures to calculate proximity values for potential technology connections.
Table 2.
The structural proximity index utilized gauges the collection of neighboring nodes and the node’s degree within a set.
Fig 5.
Illustration of technology topic networks in period 1(2013-2015).
Fig 6.
Exploration of cause-and-effect relatedness between technology topics.
Table 3.
Input features employed for training models to analyze technology convergence.
Table 4.
The number of all documents based on time interval.
Fig 7.
The distribution pattern of articles and patents each year in bio-healthcare field published from 2013 to 2021.
Fig 8.
Word count distribution in the text data after preprocessing, showing the frequency of words across the dataset.
Fig 9.
Distribution of document counts across various topics, excluding outliers.
Fig 10.
The distribution of c-TF-IDF scores across terms within topics (a) the visualization of term scores for each topic (b) the visualization of term scores with logarithmic scaling.
Fig 11.
Visualization of embedding reduction with fine-tune topic representation.
Fig 12.
Training on labeled data in pairs to forecast whether the pairs will occur in the subsequent period.
Table 5.
The number of pairs between technology topics to be used as training and test datasets.
Table 6.
Descriptive statistics data for input features in period 1.
Table 7.
The results from the performance of individual classification model and voting classifier.
Fig 13.
ROC curve for individual models and voting classifier.
Table 8.
Potential technology convergence.