Music Trend Forecast Based on LSTM

At present, accurately predicting fashion trends and predicting music black horses has become the key to attracting and catering to the masses. In view of the superiority of deep learning in data mining, data analysis and processing, this paper analyzes the music dataset, understands the distribution and characteristics of the data, and preprocesses the dataset. Aiming at the trend of the singer song playback volume in the data set within 6 months, the training set of the neural network is slid by combining the corresponding song play amount, play mean value and variance as the samples of the corresponding time points. Based on the above analysis and processing, this paper proposes a music popular trend prediction model based on LSTM deep learning, and designs and implements experiments for this model. The experimental results show that the LSTM deep learning based music trend prediction method used in this paper is more obvious than the traditional machine learning method (SVM, BP neural network, etc.) in the root mean square error (RMSE) index.