A Multimodal Music Emotion Classification Method Based on Multifeature Combined Network Classifier

Table 2

Accuracy of different audio classification models.

Classification models	Angry	Happy	Relaxed	Sad	Average
Spectrogram + CNN	0.643	0.594	0.51	0.62	0.592
Spectrogram + LSTM	0.631	0.427	0.54	0.438	0.509
Spectrogram + CNN-LSTM	0.632	0.632	0.632	0.632	0.632
Our model	0.705	0.602	0.688	0.73	0.681