A quick way to identify the “nth” friends of social media users based on spatial data mining of profiles and behaviour on a service such as Twitter is described in the International Journal of Advanced Intelligence Paradigms.
D. Gandhimathi of the Research and Development Center, Bharathiar University in Coimbatore and John Sanjeev Kumar of Thiagarajar College of Engineering in Madurai, India, explain that Twitter plays an important role in intentional social action. Thus cluster analysis of users based on likes and interests might reveal otherwise latent connections between users and so allow emergent trends to be spotted more effectively and predictions made about the behaviour and actions users might take. Such insights could be of interest to research scientists, companies and their marketing departments, not-for-profit organizations and charities, and perhaps government and law enforcement in many different contexts.
The team’s unconventional quantitative analysis hooks into the geographical metadata of each user’s Twitter updates, the geotag, where that is in place and not hidden by the user to provide even richer pickings for the data miners. The team explains that their main focus was on “recommender systems” that would engage a user’s “nth” friends in a positive manner by understanding content-based or popularity-based aspects of behaviour and social action on Twitter. The team suggests that their approach could be developed into a useful recommender algorithm. However, it is also a useful tool for community discovery and for answering questions about the large-scale clustering of users.
Their tests of the approach show it to be relatively low cost in terms of computer resources needed and to provide more accurate results when compared to other approaches.
Gandhimathi, D. and Kumar, A.J.S. (2021) ‘Prediction of Nth friends using spatial data mining in social networks’, Int. J. Advanced Intelligence Paradigms, Vol. 19, Nos. 3/4, pp.410–421.