In this research, we propose using topic modeling to estimate activation probability for predicting topic diffusion on the bibliographic networks. We utilize the supervised method to predict the propagation of a specific topic. We propose a new method to calculate activation probability for an active node and an inactive node based on the meta-path and textual information using topic modeling. Firstly, based on textual information, topic modeling is suggested to measure activation probability, namely the textual information. Secondly, combining the meta-path and textual information, we propose a new method to estimate activation probability, namely the aggregated activation probability, in which the textual information is measured by topic modeling. We conduct experiments on dissimilar topics of the bibliographic network datasets. Experimental results demonstrate that topic modeling improves the accuracy of diffusion prediction compared with term frequency–inverse document frequency.
This work is licensed under a Creative Commons Attribution 4.0 International License.