Because of the complexity data sets from real world, it is difficult to classify clearly and effectively, thus we prefer adopt fuzzy clustering approaches analyze sets. However, due variety algorithms, different number clusters will lead results. The closely related division, so how determine (k ) has become a problem. Until now, many researchers have proposed utilizing validity indexes deal wi...

A scoring rule is a device for eliciting and assessing probabilistic forecasts from an agent. When dealing with continuous outcome spaces, and absent any prior insights into the structure of the agent’s beliefs, the rule should allow for a flexible reporting interface that can accurately represent complicated, multi-modal distributions. In this paper, we provide such a scoring rule based on a n...

In recent years, the optimization of multi-objective service composition in distributed systems has become an important issue. Existing work makes a smaller set Pareto-optimal solutions to represent Pareto Front (PF). However, they do not support complex mapping quality (QoS) objective space, thus having limitations providing representative solutions. We propose enhanced differential evolution ...