A multiclass graph neural network–based framework for predicting post engagement in social media
Marco Arazzi et al.
What the paper says
In this article, we present PostGage, a new framework that identifies the most suitable publishers of target content on social media, based on the level of user engagement they are expected to generate. PostGage uses a network-based data model that considers the characteristics of publishers, as well as the degree to which the semantic content and lifespan of the posts to be published overlap with those of the posts already published. Leveraging this data model, PostGage builds a multinomial classifier based on a graph neural network (GNN). The alignment between the graph-based data model and the GNN technology allows a precise classification of publishers based on their potential to generate user engagement for target content. This feature enables a more precise definition of publishing strategies on social media, as it allows the selection of the most promising publishers for a given post. The network-based model, the post ageing mechanism and the adoption of the GNN as a multiclass classifier represent PostGage’s key contributions. We also present a series of tests performed on X that demonstrate the high quality of PostGage’s results (e.g. 92% precision in predicting high-engagement posts), establishing it as an invaluable tool for various applications, including marketing and social media management.
Evidence weight
Balanced mode · F 0.40 / M 0.15 / V 0.05 / R 0.40
| F · citation impact | 0.50 × 0.4 = 0.20 |
| M · momentum | 0.50 × 0.15 = 0.07 |
| V · venue signal | 0.50 × 0.05 = 0.03 |
| R · text relevance † | 0.50 × 0.4 = 0.20 |
† Text relevance is estimated at 0.50 on the detail page — for your query’s actual relevance score, open this paper from a search result.