Distance-based outlier queries in data streams: the novel task and algorithms