Short-Term Prediction of Available Parking Space Based on Machine Learning Approaches

Abstract

Reliable short-term prediction of available parking space (APS) is the basic theory of parking guidance information system (PGIS). Based on the Intelligent parking system at the Eastern New Town, Yinzhou District, Ningbo, China, this study collected the data of parking availability in the on-street parking areas. The variation characteristics of APS were investigated and analyzed at different spatialtemporal levels. Then the APS prediction models based on Gradient Boosting Decision Tree (GBDT) and Wavelet Neural Network (WNN) were proposed. Furthermore, an improved WNN algorithm with (WA) decomposition and Particle Swarm Optimization (PSO) were presented. The original time series was decomposed and reconstructed by wavelet analysis, and the WNN algorithm found the optimal threshold of initial weight through PSO. The result of GBDT (weekday: MSE=27.37, S MSE =0, TIME=35min, weekend: MSE=9.9, S MSE =0,TIME=35min) and WA-PSO-WNN (weekday: MSE=14.93,S MSE =1.88, TIME=160.32s, weekend: MSE=12.33, S MSE =10.23, TIME=160.95s) approximated the true value. But the prediction time of GBDT was too long to be applicable to the short-term prediction of APS in this paper. Compared with the methods of GBDT, WNN, and PSO-WNN, the WA-PSO-WNN algorithm performs much better. The average differences in MSE between WA-PSO-WNN and GBDT for weekday and weekend data are 45.45% and 58.76%, respectively, indicating that WA-PSO-WNN can increase the prediction accuracy of weekday and weekend data by an average of 45.45% and 58.76% compared with the GBDT model. Finally, the application prospects of short-term APS forecasting were also discussed in reducing cruising parking behavior, reducing illegal parking behavior and adjusting dynamic parking rates to verify the importance of APS short-term forecasting.

Toward a real‐time Smart Parking Data Management and Prediction (SPDMP) system by attributes representation learning

“…As previous research mentioned, 23,54 parking availability information is not only closely related with historical pattern, the attributes information and the variance of prediction time slots (i.e., 5 min ahead and 2 h ahead) also shows significant impact to the utility level. So, another key issues worth to review is hybridizing attention and LSTM networks for sequential data‐related research.…”

“…Combining the characteristics and skills of the neural network with the features and challenges of the parking perdition problem, several customized neural networks are designed to complete the forecasting task. Generally, current research mainly uses existing machine learning models, such as, Convolution Neural Network (CNN), 53,57 Recurrent Neural Network (RNN), 23,54,58,59 and then adopts them to address the real parking challenges. These models always include several features extraction parts: spatial features extraction, temporal features extraction, and attribute information.…”

Driver Side and Traffic Based Evaluation Model for On-Street Parking Solutions

“…This section reviews three types of parking solutions: fixed sensing [8]- [10], mobile sensing [11]- [15], and data-based modeling [16], [17], whose features are summarized in Table I.…”

“…Ye et al [17] proposed a machine learning-based approach to make a short-term estimation of available parking spaces (APS). The prediction model is based on Wavelet Neural Network.…”

Predicting vacant parking space availability zone-wisely: a hybrid deep learning approach

“…D. Chuan's study [11] showed that the XGB model outperformed the Random Forest model (RF) in terms of prediction performance and efficiency. X. Ye [12] proposed an algorithm for VPS prediction based on GBDT and Wavelet Neural Network (WNN). On this basis, an improved WNN algorithm combining Wavelet Analysis (WA) decomposition and Particle Swarm Optimization (PSO) was proposed.…”