Back to Content
The diagram begins with a core collection of W O S, leading to get paper data and D I I, leading to get patent data. Both connect to data pre-processing which includes text merging, word division, and stopwords removing. The flow moves to L D A topic identification with the paper topic and patent topic. It then proceeds to the disruptive technology topic identification based on topic similarity. The structure splits into common topic and non-common topic. Under common topic, high novelty leads to strong research intensity and weak research intensity, producing hot frontier topics and emerging frontier topics, while low novelty leads to feeble frontier topics. Under non common topic, high novelty leads to potential frontier topics and low novelty leads to feeble frontier topics.
Figure 1.
The diagram begins with a core collection of W O S, leading to get paper data and D I I, leading to get patent data. Both connect to data pre-processing which includes text merging, word division, and stopwords removing. The flow moves to L D A topic identification with the paper topic and patent topic. It then proceeds to the disruptive technology topic identification based on topic similarity. The structure splits into common topic and non-common topic. Under common topic, high novelty leads to strong research intensity and weak research intensity, producing hot frontier topics and emerging frontier topics, while low novelty leads to feeble frontier topics. Under non common topic, high novelty leads to potential frontier topics and low novelty leads to feeble frontier topics.The model of disruptive technology topic identification