Charging pile charging quantity prediction method based on user charging behavior clustering

A charging pile charging quantity prediction method of an improved LSTM-TCN model based on user charging behavior clustering comprises the following steps: collecting charging order data of a charging pile, and preprocessing the charging order data; performing K-Means clustering analysis on the charging behaviors of the charging users through the charging order data to obtain a charging behavior clustering analysis result of each type of charging users; on the basis of a clustering analysis result, Pearson correlation coefficients of the charging quantity influence factors and the charging quantity are calculated, and the influence factors with the Pearson correlation coefficients larger than a certain value are selected as a final feature set; performing min-max normalization processing on the final feature set and the historical charging amount data which are selected after processing; an improved LSTM-TCN model is established in combination with an attention mechanism; and training the improved LSTM-TCN mo