Traffic Flow Prediction and Analysis in Smart Cities Based on the WND-LSTM Model

Abstract

Aiming at the problem that the road traffic flow in intelligent city is unevenly distributed in time and space, difficult to predict, and prone to traffic congestion, combined with pattern recognition and big data mining technology, this paper proposes a research method to analyze and mine the daily travel patterns of urban vehicles. This paper proposes a WND-LSTM model, which mainly includes data preprocessing, data modelling, and model implementation, to analyze the similarity of travel patterns in seasonal changes. Combining the data mining results with the data mining results, the daily travel model of road traffic vehicles in intelligent city is established. The results of the case study showed that the WND-LSTM model outperformed ARIMA (88.48%), LR (65.79%), SVR (70.46%), KNN (68.21%), SAEs (66.95%), GRU (68.43%), and LSTM (70.41%) in MAPE, respectively, with an average accuracy improvement of 71.25% (MAPE of 0.651%).