Abstract: Highlights•An approach for processing partially labeled data by label distribution is proposed.•Fuzzy uncertainty measures are constructed for label distribution data.•A new criterion measuring the irrelevant information is defined.
Loading