What do we really need to compute the Tie Strength? An empirical study applied to Social Networks

•Analysis of literature's tie strength models, their applicability and limitations.•Proposal of domain independent predictors.•Analysis of the correlation between tie strength and social media variables.•Analysis and empirical comparison of proposed variables and tie strength models.•Comparison with state-of-the-art tie strength models from the literature.

Most of existing network-based decision-support systems, such as recommender systems, require knowing users' social context and, thus, the strength of their interactions. However, previous studies related to the usage and estimation of tie strength either assume that this parameter is given or use a computational model of their own. The amount, variety and domain specific information required to apply these models makes the reproducing and reusing of existing results extremely costly or utterly impossible. In our research, we show empirically the relative importance of different social variables for the computation of the tie strength and propose a computational model independent of the Social Networks' domain. Our experiments are based on a dataset obtained from a survey that involved more than 100 participants and comprised more than 500 social ties. The dataset is the first publicly available dataset to explicitly include tie strength measures.