In 2020, I integrated a new recommender system for thinking in InfraNodus platform that I’ve been working on since 2018.
This approach is based on direct visual feedback that enables a researcher to see how their thoughts connect and what are the gaps emerging between them. Based on those gaps and discourse network metrics, an algorithm that I wrote finds the two topical clusters that could better connected and proposes the user to think of a connection between them. Such approach helps to make ideas better connected and increase their coherency.
Alternatively, if the discourse is too interconnected, the recommender system will propose the user to develop the peripheral clusters of nodes, in order to push ideas beyond the current boundaries of the context.