An effective, efficient, and ecological transport service needs a decision support system to inform the administrations, while intelligently assisting citizens' daily transportation choices. This work presents a full-fledged Gamification Platform to develop and run sustainable mobility campaigns promoting individual green mobility behaviors. We organize geolocalized and temporal information on users' tracked trips into a rich dataset that decision-makers can exploit to analyze citizens' daily transportation choices and balance the demand and the offer of the mobility ecosystem. Through the Gamification Platform, we define three different use case scenarios - i.e., mobility within the city, home-work mobility, and home-school mobility. We also evaluate the platform according to both gameful design guidelines and persuasive technology best practices. Our findings show how long-term retention and engagement can be achieved when the gameful system's design and maintenance are assisted by the Gamification Platform, merging both administrative requirements and users' needs towards an iteratively adaptive and personalized experience.

We present approximate structure learning algorithms for Bayesian networks. We discuss the two main phases of the task: the preparation of the cache of the scores and structure optimization, both with bounded and unbounded treewidth. We improve on state-of-the-art methods that rely on an ordering-based search by sampling more effectively the space of the orders. This allows for a remarkable improvement in learning Bayesian networks from thousands of variables. We also present a thorough study of the accuracy and the running time of inference, comparing bounded-treewidth and unbounded-treewidth models.

Learning a Bayesian networks with bounded treewidth is important for reducing the complexity of the inferences. We present a novel anytime algorithm (k-MAX) method for this task, which scales up to thousands of variables. Through extensive experiments we show that it consistently yields higher-scoring structures than its competitors on complete data sets. We then consider the problem of structure learning from incomplete data sets. This can be addressed by structural EM, which however is computationally very demanding. We thus adopt the novel k-MAX algorithm in the maximization step of structural EM, obtaining an efficient computation of the expected sufficient statistics. We test the resulting structural EM method on the task of imputing missing data, comparing it against the state-of-the-art approach based on random forests. Our approach achieves the same imputation accuracy of the competitors, but in about one tenth of the time. Furthermore we show that it has worst-case complexity linear in the input size, and that it is easily parallelizable.