Bus load prediction method and system based on multi-scale time sequence convolutional neural network

• Main Authors: LI HONG, ZHANG QIPEI, WANG CHAOHUI, JI WENLU, DENG XING, ZHU HONG, WANG WEI, LIU JUNJUN, WU LIN, KUANG WENTENG, LU JIXIANG, XIE FENG, SHEN MAOYA, CHEN YUCHEN
• Format: Patent
• Language: Chinese; English
• Published: 26.09.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention discloses a bus load prediction method based on a multi-scale time sequence convolutional neural network, and the method comprises the steps: obtaining historical load data of a plurality of buses and a corresponding historical weather data set, carrying out the primary processing, and extracting related features as a training set and a verification set; determining a training data set according to the correlation degree of the correlation features in the training set and the bus load prediction; respectively inputting the training data set into a one-way LSTM model, a dense link TCN model and a multi-scale CNN model for training; respectively verifying the three trained models through a verification set, determining a likelihood function coefficient according to the prediction accuracy of the three models, and constructing a fusion prediction model; and predicting the bus load through the fusion prediction model. According to the method, the distribution characteristics of the time series data