Ya Gao’s research while affiliated with East China Normal University and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (5)


Network-Scale Traffic Modeling and Forecasting with Graphical Lasso and Neural Networks
  • Preprint

December 2017

Shiliang Sun

·

Rongqing Huang

·

Ya Gao

Traffic flow forecasting, especially the short-term case, is an important topic in intelligent transportation systems (ITS). This paper does a lot of research on network-scale modeling and forecasting of short-term traffic flows. Firstly, we propose the concepts of single-link and multi-link models of traffic flow forecasting. Secondly, we construct four prediction models by combining the two models with single-task learning and multi-task learning. The combination of the multi-link model and multi-task learning not only improves the experimental efficiency but also the prediction accuracy. Moreover, a new multi-link single-task approach that combines graphical lasso (GL) with neural network (NN) is proposed. GL provides a general methodology for solving problems involving lots of variables. Using L1 regularization, GL builds a sparse graphical model making use of the sparse inverse covariance matrix. In addition, Gaussian process regression (GPR) is a classic regression algorithm in Bayesian machine learning. Although there is wide research on GPR, there are few applications of GPR in traffic flow forecasting. In this paper, we apply GPR to traffic flow forecasting and show its potential. Through sufficient experiments, we compare all of the proposed approaches and make an overall assessment at last.


Network-Scale Traffic Modeling and Forecasting with Graphical Lasso and Neural Networks

November 2012

·

167 Reads

·

129 Citations

Journal of Transportation Engineering

Traffic flow forecasting, especially the short-term case, is an important topic in intelligent transportation systems (ITS). This paper researches network-scale modeling and forecasting of short-term traffic flows. First, the concepts of single-link and multilink models of traffic flow forecasting are proposed. Secondly, four prediction models are constructed by combining the two models with single-task learning (STL) and multitask learning (MTL). The combination of the multilink model and multitask learning not only improves the experimental efficiency but also improves the prediction accuracy. Moreover, a new multilink, single-task approach that combines graphical lasso (GL) with neural network (NN) is proposed. GL provides a general methodology for solving problems involving lots of variables. Using L1 regularization, GL builds a sparse graphical model, making use of the sparse inverse covariance matrix. Gaussian process regression (GPR) is a classic regression algorithm in Bayesian machine learning. Although there is wide research on GPR, there are few applications of GPR in traffic flow forecasting. In this paper, GPR is applied to traffic flow forecasting, and its potential is shown. Through sufficient experiments, all of the proposed approaches are compared, and an overall assessment is made.


Network-Scale Traffic Modeling and Forecasting with Graphical Lasso

May 2011

·

129 Reads

·

26 Citations

Lecture Notes in Computer Science

Traffic flow forecasting is an important application domain of machine learning. How to use the information provided by adjacent links more efficiently is a key to improving the performance of Intelligent Transportation Systems (ITS). In this paper, we build a sparse graphical model for multi-link traffic flow through the Graphical Lasso (GL) algorithm and then implement the forecasting with Neural Networks. Through a large number of experiments, we find that network-scale traffic forecasting with modeling by Graphical Lasso performs much better than previous research. Traditional approaches considered the information provided by adjacent links but did not extract the information. Thus, although they improved the performance to some extent, they did not make good use of the information. Furthermore, we summarize the theoretical analysis of Graphical Lasso algorithm. From theoretical and practical points of view, we fully verify the superiority of Graphical Lasso used in modeling for multi-link traffic flow forecasting.


Multi-link traffic flow forecasting using neural networks

August 2010

·

30 Reads

·

20 Citations

Traffic flow forecasting is an important application of computational intelligence and an active research topic in Intelligent Transportation Systems (ITS). However, traditional methods called single-link traffic flow forecasting usually predict one link's unidirectional traffic flow at a time, which do not take the relevance of adjacent links into account and make the ITS have a low efficiency. In this paper, we propose a new approach named multi-link traffic flow forecasting using neural networks (NNs), which can predict traffic flows on all the road links of one junction simultaneously. Experimental results show that it can not only increase the efficiency of ITS but also improve the performance of prediction. Furthermore, we combine multi-task learning with the multi-link traffic flow forecasting and obtain a better performance of prediction. All these experiments indicate that the multi-link traffic flow forecasting is a much more effective approach for traffic flow forecasting.


An empirical evaluation of linear and nonlinear kernels for text classification using Support Vector Machines

August 2010

·

84 Reads

·

27 Citations

This paper compares the performance of linear and nonlinear kernels of Support Vector Machines (SVM) used for text classification. The study is motivated by the previous viewpoint that linear SVM performs better than nonlinear one, and that, although there are many investigations have proved that SVM performs well in text classification, there is no serious investigation on the comparison between linear SVM and nonlinear SVM. In our study, we carry out two experiments with different datasets and use grid-search on the selection of kernel parameters. Empirical results show that, in fact, nonlinear SVM performs better than linear SVM as long as with appropriate kernel parameters. This conclusion will provide useful guidance for people applying SVM to text classification and other corresponding fields.

Citations (4)


... [37] introduces a space-time diurnal (ST-D) method in which link-wise travel time correlation at multiple lag time is utilized. [32] utilizes Gaussian process regression (GPR) model and graphic Lasso to forecast traffic flow. 3 [2]proposes a KNN model to forecast travel time up to one hour ahead, the model uses redefined inter-segments distances by incorporating the grade of connectivity between road segments, and considers spatial-temporal correlations and state matrices to identify traffic state. ...

Reference:

Interpretable mixture of experts for time series prediction under recurrent and non-recurrent conditions
Network-Scale Traffic Modeling and Forecasting with Graphical Lasso and Neural Networks
  • Citing Article
  • November 2012

Journal of Transportation Engineering

... Park et al. [35] used feedforward multilayer neural networks for estimating link travel times on freeways. Other applications of ANNs to short-term forecasting can be found in [36][37][38][39][40]. Ledoux [41] proposed the use of ANNs within an urban traffic flow model, while Florio and Mussone [42] studied traffic flow stability on freeways with neural network models. ...

Multi-link traffic flow forecasting using neural networks
  • Citing Conference Paper
  • August 2010

... The input data was mapped into a higher-dimensional feature space by the nonlinear sigmoid kernel, where it was linearly separable. Text classification tasks like sentiment analysis and topic classification were accomplished using SVM with the sigmoid kernel [9]. The sigmoid kernel's nonlinear properties enabled complex correlations between words or attributes in textual input to be captured. ...

An empirical evaluation of linear and nonlinear kernels for text classification using Support Vector Machines
  • Citing Conference Paper
  • August 2010

... Transport and mobility are core components that support the operation and function of modern cities, and they evolve with the development of urban society. Nowadays, urban areas worldwide are expanding rapidly with high complexity (US Census Bureau, 2011;Wang et al., 2018;Guo et al., 2019), posing serious challenges to society (Gao et al., 2011;Stefaniec et al., 2020). Transport sustainability has long been a concern in virtually all large cities worldwide (Alonso et al., 2015). ...

Network-Scale Traffic Modeling and Forecasting with Graphical Lasso
  • Citing Conference Paper
  • May 2011

Lecture Notes in Computer Science