Prediction intervals to account for uncertainties in neural network predictions: Methodology and application in bus travel time prediction

Mazloumi, Ehsan; Rose, Geoff; Currie, Graham; Moridpour, Sara

doi:10.1016/j.engappai.2010.11.004

Cited by 120 publications

(77 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The performance of a NN often deteriorates when the number of input variables increases. This has been referred to as the curse of dimensionality in the literature [26]. Increasing the number of input variables also leads to the need to use more training examples and process time to effectively understand the input-output relationship.…”

Section: Data Analysis and Training Of The Networkmentioning

confidence: 99%

“…In this process, a NN constructs an input-output mapping and adjusts the weights and the biases at each iteration based on the minimization of an error measure between the produced and the desired outputs. The adequate selection of inputs, hidden layers, training functions, tapped delay lines, and number of neurons strongly influences the success of the training process [26,27].…”

Section: Forecasting With Neural Networkmentioning

confidence: 99%

See 1 more Smart Citation

RFID card security for public transportation applications based on a novel neural network analysis of cardholder behavior characteristics

DÜZENLİ¹

2015

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

This paper proposes a novel approach that applies neural network forecasting to security for closed-loop prepaid cards based on low-cost technologies such as RFID and 1-Wire. The security vulnerability of low-cost RFID closed-loop prepaid card systems originates mostly from the card itself. Criminal organizations counterfeit or clone card data. Although high-security prepaid cards exist, they are often too expensive for transport ticketing, and even their security is not guaranteed for a well-defined period of time. Therefore, data encryption systems are used widely against counterfeiting with success. However, it has not been possible to develop countermeasures with comparable success against cloning. Our proposed security application uses neural network forecasting to determine the recharge day behavioral characteristics of the cardholder and predict the next time the cardholder will recharge their card. Based on the prediction for the recharge time, the expiration date of the low-cost RFID prepaid card is defined, which is a good countermeasure against cloning. FTDNN, LRNN, and NARX network architectures with one hidden layer are considered in this research. The effects of the network architecture, the number of neurons, the training algorithm, and the prediction performance function on the recharge day forecast are investigated. Experimental results confirm the accuracy of the recharge time forecast and confirm countermeasures against cloning. Our proposed security approach with neural network forecasting is applied with success to the Turkish public transport without an online backend system.

show abstract

Section: Data Analysis and Training Of The Networkmentioning

confidence: 99%

Section: Forecasting With Neural Networkmentioning

confidence: 99%

RFID card security for public transportation applications based on a novel neural network analysis of cardholder behavior characteristics

DÜZENLİ¹

2015

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

show abstract

“…Most of the traditional prediction models belong to the first category, including historical average and smoothing techniques, parametric and non-parametric regression [22][23][24], autoregressive integrated moving average (ARIMA) [25][26][27], machine learning [28], fuzzy logic [29,30] and neural networks [31][32][33]. These methods often suffer from high computational complexity either due to the stationery requirements or a large number of estimated parameters and may not be adaptive to the change in traffic patterns [34].…”

Section: Traffic Predictionmentioning

confidence: 99%

Real-time urban traffic amount prediction models for dynamic route guidance systems

Liang

Wakahara

2014

J Wireless Com Network

View full text Add to dashboard Cite

The route guidance system (RGS) has been considered an important technology to mitigate urban traffic congestion. However, existing RGSs provide only route guidance after congestion happens. This reactive strategy imposes a strong limitation on the potential contribution of current RGS to the performance improvement of a traffic network. Thus, a proactive RGS based on congestion prediction is considered essential to improve the effectiveness of RGS. The problem of congestion prediction is translated into traffic amount (i.e. the number of vehicles on the individual roads) prediction, as the latter is a straightforward indicator of the former. We thereby propose two urban traffic prediction models using different modeling approaches. Model-1 is based on the traffic flow propagation in the network, while Model-2 is based on the time-varied spare flow capacity on the concerned road links. These two models are then applied to construct a centralized proactive RGS. Evaluation results show that (1) both of the proposed models reduce the prediction error up to 52% and 30% in the best cases compared to the existing Shift Model, (2) providing proactive route guidance helps reduce average travel time by up to 70% compared to providing reactive one and (3) non-rerouted vehicles could benefit more from route guidance than rerouted vehicles do.

show abstract

“…Depending on the goal and scope of the task, classification or regression can be an intermediary step in the overall process of knowledge extraction. Mazloumi et al (2011) used neural networks to learn from travel time data in order to build prediction models, but then computed uncertainty in the model taking into account different sources of error, so that a prediction interval was obtained instead of a single point estimate. Varga et al (2009) applied decision trees to reactor run-away data to organize and summarize the most important information from the process and then display it in a userfriendly format in an operator support system.…”

Section: Introductionmentioning

confidence: 99%

Conditional inference trees for knowledge extraction from motor health condition data

Sard-Espinosa

Subbiah

Bartz-Beielstein

2017

Engineering Applications of Artificial Intelligence

View full text Add to dashboard Cite

As the amount of data gathered by monitoring systems increases, using computational tools to analyze it becomes a necessity. Machine learning algorithms can be used in both regression and classification problems, providing useful insights while avoiding the bias and proneness to errors of humans. In this paper, a specific kind of decision tree algorithm, called conditional inference tree, is used to extract relevant knowledge from data that pertains to electrical motors. The model is chosen due to its flexibility, strong statistical foundation, as well as great capabilities to generalize and cope with problems in the data. The obtained knowledge is organized in a structured way and then analyzed in the context of health condition monitoring. The final results illustrate how the approach can be used to gain insight into the system and present the results in an understandable, user-friendly manner.

show abstract

Prediction intervals to account for uncertainties in neural network predictions: Methodology and application in bus travel time prediction

Cited by 120 publications

References 34 publications

RFID card security for public transportation applications based on a novel neural network analysis of cardholder behavior characteristics

RFID card security for public transportation applications based on a novel neural network analysis of cardholder behavior characteristics

Real-time urban traffic amount prediction models for dynamic route guidance systems

Conditional inference trees for knowledge extraction from motor health condition data

Contact Info

Product

Resources

About