Music Trend Prediction Based on Improved LSTM and Random Forest Algorithm

As one of the entertainment consumption products, pop music attracts more and people’s attention. In context big data, many listeners can determine development trend to a large extent. order predict music, we dig analyze audience’s preferences deeply based on massive user data. This paper proposes prediction method improved LSTM random forest algorithm. The algorithm first performs abnormal data processing normalization test set. Then important features are selected by corrected rough set compensation system. Finally, is made improving LSTM. experiment, RMSE MAER used as performance evaluation indexes algorithm, results show that proposed better popularity trend. At same time, root means square error mean absolute index obviously.