New Method of Song Sentiment Classification Based on Neural Network
Ying Yu · 2022
In order to classify song text information more effectively, this paper proposes a new neural network classification model based on emotion vector. On the basis of Word2vec, a double word-embedding word vector containing emotion vector is constructed, and the vector is used as the input of Long Short Term Memory (LSTM). LSTM neural network is trained to extract the features of song text and realize emotional classification of song text. The experimental results show that the method can better embody the emotional features of text, and has higher accuracy than the traditional LSTM text classification.