Text this: Individual Contribution-Based Spatial-Temporal Attention on Skeleton Sequences for Human Interaction Recognition