This study addresses swarm intelligence-based approaches in data quality detection. First, three typical swarm intelligence models and their applications in abnormity detection are introduced, including Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), Bee Colony Optimization (BCO). Then, it presents three approaches based on ACO, PSO and BCO for detection of attribute outliers in datasets. These approaches use different search strategies on the data items; however, they choose the same fitness function (i.e. the O-measure) to evaluate the solutions, and they make use of swarms of the fittest agents and random moving agents to obtain superior solutions by changing the searching paths or positions of agents. Three algorithms are described and explained, which are efficient by heuristic principles.
展开▼
