1,
610000
–>
610500

2, Label distribution learning(LDL)
–>Label distribution learning (LDL)

3,
Howerver, existing LDL algorithms usually extract sample features from a single view, and the features extracted from a single view often cannot completely describe the sample.
–>
Existing LDL algorithms usually extract sample features from a single view.
However, a single view cannot perform rich feature representation, resulting in the training model not being stable enough.

Multi-view can describe the sample in many different ways, and get more information about the sample.
In this paper, we proposed a multi-view-based label learning algorithm named MSN-LDL, which uses Multiple Sub-Networks to extract the semantic features of sample from different views.
–>
In this paper, we propose a multi-view-based label learning algorithm named MSN-LDL, which uses Multiple Sub-Networks to extract the semantic features of sample from different views.