K-means clustering for the aggregation of HFLTS possibility distributions: N-two-stage algorithmic paradigm.