Research on Music Emotion Classification Based on the Feature Model of Emotion Category Contribution

There are problems in the text vector space model: the high dimension causes the sparsity of the vector space, the classification contribution is lower, and the classification accuracy, etc. The characteristic model for the contribution of special words to the emotion category is presented, An improved classification dimension decision fusion method based on the combination of audio features and lyrics features, Experimental results show that this method has good classification effect.