This work is funded by Umeå University in Sweden on federated database research.

Purpose Road accidents have come to be considered a major public health problem worldwide. The aim of many studies is therefore to identify the main factors contributing to the severity of crashes. Methods This paper examines a large-scale data mining technique known as association rule mining, which can predict future accidents in advance and allow drivers to avoid the dangers. However, this technique produces a very large number of decision rules, preventing decision makers from making their own selection of the most relevant rules. In this context, the integration of a multi-criteria decision analysis approach would be particularly useful for decision makers affected by the redundancy of the extracted rules. Conclusion An analysis of road accidents in the province of Marrakech (Morocco) between 2004 and 2014 shows that the proposed approach serves this purpose; it may provide meaningful information that could help in developing suitable prevention policies to improve road safety.

No abstract

Federated machine learning has great promise to overcome the input privacy challenge in machine learning. The appearance of several projects capable of simulating federated learning has led to a corresponding rapid progress on algorithmic aspects of the problem. However, there is still a lack of federated machine learning frameworks that focus on fundamental aspects such as scalability, robustness, security, and performance in a geographically distributed setting. To bridge this gap we have designed and developed the FEDn framework. A main feature of FEDn is to support both cross-device and cross-silo training settings. This makes FEDn a powerful tool for researching a wide range of machine learning applications in a realistic setting.

Recently, association rule mining plays a vital role in knowledge discovery in database. In fact, in most cases, the real datasets lead to a very large number of rules, which do not allow users to make their own selection of the most relevant. The difficult task is mining useful and nonredundant rules. Several approaches have been proposed, such as rule clustering, informative cover method and quality measurements. Another way to selecting relevant association rules, we believe that it is necessary to integrate a decisional approach within the knowledge discovery process. Therefore, in this paper, we propose an approach to discover a category of relevant association rules based on multi-criteria analysis. In other side, the general process of association rules extraction becomes more and more complex, to solve such problem, we also proposed a multi-agent system for modeling the different process of our proposed approach. Therefore, we conclude our work by an empirical study applied to a set of banking data to illustrate the performance of our approach.