Action extraction from social networks

摘要

Data mining methods focus on discovering models and patterns from large databases that summarize the data. However, generating such results is not an end in itself because their applicability is not straightforward. Ideally, the user would ultimately like to use them to decide what actions to take. Action mining explicitly emerged as a response to this need. Currently, most of the action mining methods rely on simple data which describes each object independently that means they do not take into account relationships between objects. In social networks, relationships enable an individual to influence another one, so ignoring them in action mining process would lead to miss some profitable actions. In this paper, we introduce action mining from social networks. In fact, our main contribution is to extract cost-effective actions which is formulated as an optimization problem where the objective is to learn actions consisting of the changes in the network that are likely to result in desired changes in the labels of intended individuals while minimizing the cost of the changes. Experiments confirm that the proposed approach performs much better than the current state-of-the-art in action mining.