Towards a General Prediction System for the Primary Delay in Urban Railways

Abstract

Nowadays a large amount of data is collected from sensor devices across the cyber-physical networks. Accurate and reliable primary delay predictions are essential for rail operations management and planning. However, very few existing 'big data' methods meet the specific needs in railways. We propose a comprehensive and general data-driven Primary Delay Prediction System (PDPS) framework, which combines General Transit Feed Specification (GTFS), Critical Point Search (CPS), and deep learning models to leverage the data fusion. Based on this framework, we have also developed an open source data collection and processing tool that reduces the barrier to the use of the different open data sources. Finally, we demonstrate an advanced deep learning model, the novel ConvLSTM Encoder-Decoder model with CPS for better primary delay predictions.

Towards A More Effective Bidirectional LSTM-Based Learning Model for Human-Bacterium Protein-Protein Interactions

“…Bidirectional LSTM model (Bi-LSTM) is the critical component of the model, which is a variant deep learning model of LSTM proposed by [18,30]. LSTM model and its variant version Bi-LSTM have demonstrated superior performance in domainss such as natural language processing , transportation and action recognition [36,38]. In Bi-LSTM model, two layers, namely forward and backward layers, are designed to converge into a single layer.…”

Big data processing and analysis on the impact of COVID-19 on public transport delay

“…Nowadays GTFS has been used as an industry standard for a majority of transit agencies to publish their transit data around the world [ 6 ]. As GTFS data contains both scheduled and real-time information about transit operations, it has been actively used for many research problems such as transit accessibility [ [7] , [8] , [9] , [10] , [11] ], transit network analysis [ 12 , 13 ], performance evaluation [ 14 , 15 ], delay prediction [ [16] , [17] , [18] ], and transit trip inference [ 19 , 20 ].…”

A Hybrid LSTM-CPS Approach for Long-Term Prediction of Train Delays in Multivariate Time Series

“…A classification algorithm is required to incorporate domain knowledge to extract the information from the output of a predictive model. CPS is a rule-based classification method to search primary, secondary and on-time points [28]. It can be extended to identify those train stations that cause major disruption to services.…”

“…Ref. [28] proposed a generic delay prediction framework to establish a connection between expert knowledge with long short-term memory (LSTM). The Convolutional Neural Network (CNN) [29], Gated Recurrent Unit (GRU) [30], and LSTM can be combined to form a more complex structure.…”