The Ensemble Approach to Forecasting: A Review and Synthesis

Recently published:

  • Wu, Hao, and Levinson, D. (2021) The Ensemble Approach to Forecasting: A Review and Synthesis. Transportation Research part C. Volume 132, 103357 [doi] [Author Link]
  • Highlights

    • Review and synthesize methods of ensemble forecasting with a unifying framework.
    • As decision support tools, ensemble models systematically account for uncertainties.
    • Ensemble methods can include combining models, data, and ensemble of ensembles.
    • Transport ensemble models have the potential for improving accuracy and reliability.

    Abstract

    Ensemble forecasting is a modeling approach that combines data sources, models of different types, with alternative assumptions, using distinct pattern recognition methods. The aim is to use all available information in predictions, without the limiting and arbitrary choices and dependencies resulting from a single statistical or machine learning approach or a single functional form, or results from a limited data source. Uncertainties are systematically accounted for. Outputs of ensemble models can be presented as a range of possibilities, to indicate the amount of uncertainty in modeling. We review methods and applications of ensemble models both within and outside of transport research. The review finds that ensemble forecasting generally improves forecast accuracy, robustness in many fields, particularly in weather forecasting where the method originated. We note that ensemble methods are highly siloed across different disciplines, and both the knowledge and application of ensemble forecasting are lacking in transport. In this paper we review and synthesize methods of ensemble forecasting with a unifying framework, categorizing ensemble methods into two broad and not mutually exclusive categories, namely combining models, and combining data; this framework further extends to ensembles of ensembles. We apply ensemble forecasting to transport related cases, which shows the potential of ensemble models in improving forecast accuracy and reliability. This paper sheds light on the apparatus of ensemble forecasting, which we hope contributes to the better understanding and wider adoption of ensemble models.

    Fig. 1. Methods of combining data and models.

    Making Accessibility Work in Practice

    Recently Published:

    • El-Geneidy, Ahmed and Levinson, D. (2021) Making Accessibility Work in Practice. Transport Reviews [doi] [first 50 free download]

    Accessibility, the ease of reaching destination, is the most comprehensive land use and transport systems performance measure (Levinson & Wu, 2020; Wachs & Kumagai, 1973; Wu & Levinson, 2020). Accessibility has been applied in planning research since the 1950s (Hansen, 1959), and still today, we find major barriers to adopting it in practice (Handy, 2020). Advances in computing and software have enabled researchers to generate complex measures of accessibility with higher spatial and temporal resolutions moving accessibility research at a fast pace, while the implementation of accessibility, in practice, lags (Boisjoly & El-Geneidy, 2017). Even simple measures, such as the cumulative opportunities measures of accessibility, confront challenges in adoption.

    Immigrant Settlement Patterns, Transit Accessibility, and Transit Use

    Recently published:

  • Allen, Jeff, Farber, Steven, Greaves, Stephen, Clifton, Geoffrey, Wu, Hao, Sarkar, Hao, and Levinson, D. (2021) Immigrant Settlement Patterns, Transit Accessibility, and Transit Use. Journal of Transport Geography. 96 103187 [doi]
  • Abstract: Public transit is immensely important among recent immigrants for enabling daily travel and activity participation. The objectives of this study are to examine whether immigrants settle in areas of high or low transit accessibility and how this affects transit mode share. This is analyzed via a novel comparison of two gateway cities: Sydney, Australia and Toronto, Canada. We find that in both cities, recent immigrants have greater levels of public transit accessibility to jobs, on average, than the overall population, but the geography of immigrant settlement is more suburbanized and less clustered around commuter rail in Toronto than in Sydney. Using logistic regression models with spatial filters, we find significant positive relationships between immigrant settlement patterns and transit mode share for commuting trips, after controlling for transit accessibility and other socio-economic factors, indicating an increased reliance on public transit by recent immigrants. Importantly, via a sensitivity analysis, we find that these effects are greatest in peripheral suburbs and rural areas, indicating that recent immigrants in these areas have more risks of transport-related social exclusion due to reliance on insufficient transit service.

    Fig. 3. Bivariate maps of transit accessibility and density of recent immigrants.

    Towards a General Theory of Access: Video

    Levinson, D. M., & Wu, H. (2020). Towards a general theory of access. Journal of Transport and Land Use, 13(1), 129-158. https://doi.org/10.5198/jtlu.2020.1660

    This paper integrates and extends many of the concepts of accessibility deriving from Hansen’s (1959) seminal paper, and develops a theory of access that generalizes from the particular measures of access that have become increasingly common. Access is now measured for a particular place by a particular mode for a particular purpose at a particular time in a particular year. General access is derived as a theoretical ideal that would be measured for all places, all modes, all purposes, at all times, over the lifecycle of a project. It is posited that more general access measures better explain spatial location phenomena.

    Job and Worker Density and Transit Network Dynamics

    Recently published:

    • Li, Manman, Cui, Mengying, and Levinson, D. (2021) Job and Worker Density and Transit Network Dynamics. International Journal of Sustainable Transportation. [doi]

    This paper proposes a general framework to explore the interaction between land use and transport systems. Hypotheses about those relationships are generated. A series of statistical tests are conducted to explain the co-development of land use and transit networks for metropolitan areas at a micro-geographic scale and to disentangle causes and effects. The specific case of Minneapolis – Saint Paul (Twin Cities) metropolitan is examined using a panel of block-level land use and stop-level transit data. The results show that the development of land use, specifically, resident workers, can lead to the increase in bus demand, and thus further induce the increase in bus supply; the co-development of bus demand and supply is simultaneous on a yearly basis.

    Hypotheses about relationships between land use and bus network

    Urban access across the globe: an international comparison of different transport modes

    Recently published

    • Hao Wu, Paolo Avner, Genevieve Boisjoly, Carlos K. V. Braga, Ahmed El-Geneidy, Jie Huang, Tamara Kerzhner, Brendan Murphy, Michał A. Niedzielski, Rafael H. M. Pereira, John P. Pritchard, Anson Stewart, Jiaoe Wang, and David Levinson (2021) Urban access across the globe: an international comparison of different transport modes. NPJ Urban Sustainability Vol. 1, Article 16 [doi]

    Access (the ease of reaching valued destinations) is underpinned by land use and transport infrastructure. The importance of access in transport, sustainability, and urban economics is increasingly recognized. In particular, access provides a universal unit of measurement to examine cities for the efficiency of transport and land use systems. This paper examines the relationship between population-weighted access and metropolitan population in global metropolitan areas (cities) using 30-minute cumulative access to jobs for 4 different modes of transport; 117 cities from 16 countries and 6 continents are included. Sprawling development with intensive road network in American cities produces modest automobile access relative to their sizes, but American cities lag behind globally in transit and walking access; Australian and Canadian cities have lower automobile access, but better transit access than American cities; combining compact development with an intensive network produces the highest access in Chinese and European cities for their sizes. Hence density and mobility co-produce better access. This paper finds access to jobs increases with populations sublinearly, so doubling metropolitan population results in a less than double access to jobs. The relationship between population and access characterizes regions, countries and cities, and significant similarities exist between cities from the same country.

    The Perception of Access in Sydney

    Recently published:

    Based on a survey of 197 Sydneysiders, this study shows residents overestimated the attractiveness of the city centre compared to the entire metropolitan area, as well as the number of jobs they can reach from home. They also overestimated travel times compared to Google Maps, especially for travel times by car.

    The Economics of Findings

    Recently published:

    Abstract

    This paper considers the monetary and time costs of producing Findings (formerly Transport Findings). After enumerating the journal’s expenses, we find the marginal monetary cost of an article is, on average, about $65, and that the journal incurs $1966 in fixed costs per year. Also, using data from a survey of Findings’ reviewers and estimate of reviewers’ value of time, we also calculate the time costs of operating findings. Most reviewers agree that compensating them for producing timely reviews would be an effective inventive.

    Injury severity prediction from two-vehicle crash mechanisms with machine learning and ensemble models

    • Ji, Ang and Levinson, D. (2020) Injury severity prediction from two-vehicle crash mechanisms with machine learning and ensemble models. IEEE Open Journal of Intelligent Transportation Systems. [doi]

    Machine learning algorithms aim to improve the power of predictors over conventional regression models. This study aims to tap the predictive potential of crash mechanism- related variables using ensemble machine learning models. The results demonstrate selected models can predict severity at a high level of accuracy. The stacking model with a linear blender is preferred for the designed ensemble combination. Most bagging, boosting, and stacking algorithms perform well, indicating en- semble models are capable of improving upon individual models.