Real-Time Human-Music Emotional Interaction Based on Deep Learning and Multimodal Sentiment Analysis

Table 3

Accuracy of corresponding to different weights.

Weight of text	Weight of image	Accuracy
0.1	0.9	0.8179
0.3	0.7	0.8421
0.5	0.5	0.8539
0.7	0.3	0.8727
0.9	0.1	0.8635