Applying gradient boosting decision trees to examine non-linear effects of the built environment on driving distance in Oslo

Ding, Chuan; Cao, Xinyu; Næss, Petter

doi:10.1016/j.tra.2018.02.009

Cited by 186 publications

(128 citation statements)

References 42 publications

Supporting

Mentioning

121

Contrasting

Unclassified

Order By: Relevance

“…The function of

. To minimize the loss function, a residual term can be added to the prediction function F ( x ) in each iteration [ 30 , 31 ]. The steps of this algorithm can be concluded as follows: Initialize model as constant c :

For

, as follows: For

, compute the so-called pseudo-residuals, as follows:

Fit decision tree

by training set

;

Calculate weights

by minimizing the loss function

Update

; Output

…”

Section: Methodsmentioning

confidence: 99%

Identify Risk Pattern of E-Bike Riders in China Based on Machine Learning Framework

Wang

Kou

Song

2019

Entropy

View full text Add to dashboard Cite

In this paper, the risk pattern of e-bike riders in China was examined, based on tree-structured machine learning techniques. Three-year crash/violation data were acquired from the Kunshan traffic police department, China. Firstly, high-risk (HR) electric bicycle (e-bike) riders were defined as those with at-fault crash involvement, while others (i.e. non-at-fault or without crash involvement) were considered as non-high-risk (NHR) riders, based on quasi-induced exposure theory. Then, for e-bike riders, their demographics and previous violation-related features were developed based on the crash/violation records. After that, a systematic machine learning (ML) framework was proposed so as to capture the complex risk patterns of those e-bike riders. An ensemble sampling method was selected to deal with the imbalanced datasets. Four tree-structured machine learning methods were compared, and a gradient boost decision tree (GBDT) appeared to be the best. The feature importance and partial dependence were further examined. Interesting findings include the following: (1) tree-structured ML models are able to capture complex risk patterns and interpret them properly; (2) spatial-temporal violation features were found as important indicators of high-risk e-bike riders; and (3) violation behavior features appeared to be more effective than violation punishment-related features, in terms of identifying high-risk e-bike riders. In general, the proposed ML framework is able to identify the complex crash risk pattern of e-bike riders. This paper provides useful insights for policy-makers and traffic practitioners regarding e-bike safety improvement in China.

show abstract

“…The function of

For

, as follows: For

, compute the so-called pseudo-residuals, as follows:

Fit decision tree

by training set

;

Calculate weights

by minimizing the loss function

Update

; Output

…”

Section: Methodsmentioning

confidence: 99%

Identify Risk Pattern of E-Bike Riders in China Based on Machine Learning Framework

Wang

Kou

Song

2019

Entropy

View full text Add to dashboard Cite

show abstract

“…Obviously, in areas with many local amenities, the variation is greater and a certain category of amenity is more likely to be represented. (2) Previous research shows that the effect of the built environment is often non-linear (Ellde´r, 2018;Ding et al, 2018;Tao et al, 2020;van Wee and Handy, 2016) and varies between contexts (Naess, 2019;Salon, 2015;Voulgaris et al, 2016). Acknowledging this means that the effect of establishing or closing amenities will depend on the other types of amenities present.…”

Section: Literature Reviewmentioning

confidence: 99%

“…From a theoretical perspective, previous studies generally assume that the effect of the built environment on travel is linear, that is, that a change in supply density will generate a proportional change in car travel, regardless of context. Recent studies suggest the existence of non-linear effects (Case, 2013;Ding et al, 2018;Tao et al, 2020) and that effects vary amongst different geographical areas (Boarnet et al, 2011;Salon, 2009Salon, , 2015. However, these studies mainly focus on non-linearity/heterogeneity regarding aggregate 'D variables'.…”

Section: Introductionmentioning

confidence: 99%

When local access matters: A detailed analysis of place, neighbourhood amenities and travel choice

2020

View full text Add to dashboard Cite

How cities can become more conducive to sustainable travel patterns is a pressing matter in urban research. In the built environment–travel literature, several features potentially promoting sustainable mobility remain under-researched. One concerns the precise role of local accessibility – in terms of the number, variety and specific types of amenities – in promoting the slow modes of walking and cycling and in reducing car use. Another is the possibility that the standard assumption of a linear relationship between density and (un)sustainable travel is flawed. To fill these gaps, we explore specific local amenity supply types that stimulate a large share of residents to walk or cycle, avoiding long drives, to accomplish their daily activities. In a case study of the Västra Götaland region in Sweden, we analyse a combination of high-quality micro data from registers and the national travel survey using a stepwise statistical approach. We find evidence of non-linear discontinuities and significant geographical variation in how local accessibility affects travel behaviour. While more neighbourhood amenities contribute to more walking/cycling, the effect of establishing additional amenities varies amongst neighbourhood types and travel variables. Local everyday services, for example, grocery stores and preschools, are critical in triggering change. The findings inform urban planning by showing how a basic supply of essential amenities could contribute to better prerequisites for walking or cycling and reduced car dependence by inducing residents to choose local options.

show abstract

“…The GBDT algorithm has been demonstrated to be an efficient, high-precision, low-bias model in many practices and has been widely used by data scientists in tasks, such as classification and regression. A complete mathematical and technical description of the GBDT model can be found in [19] and [62].…”

Section: Gbdt Modeling Methodsmentioning

confidence: 99%

Real-Time Prediction of the Water Accumulation Process of Urban Stormy Accumulation Points Based on Deep Learning

Zhou

Wang

2020

IEEE Access

View full text Add to dashboard Cite

Influenced by climate change and urbanization, urban flood frequently occurs and represents a serious challenge for many cities. Therefore, it is necessary to generate refined predictions of urban floods, such as the prediction of water accumulation processes at water accumulation points, which is of great significance for supporting water-related managers to reduce flood losses. In this study, 16 combination schemes of rainfall sensitivity indicators were used to determine the optimal scheme for predicting the depth of accumulated water, and the gradient boosting decision tree (GBDT) algorithm in deep learning was used to build a prediction model of the accumulation process of urban stormy accumulation points. Among the 16 schemes, the relative error of scheme 1 is 15.39%, and the qualified rate is 92.86%. This scheme exhibits the highest accuracy for the prediction results of water accumulation depth. Given this finding, the GBDT algorithm was used to construct a regression prediction model of the water accumulation process based on the collected historical rainfall water accumulation data of 50 water accumulation points. The results demonstrated that the GBDT regression prediction model has a mean relative error of 19.77%, a qualified rate of 82.00%, and a peak average relative error of 5.48%, which verify the validity and applicability of the model for the real-time prediction of the process of water accumulation. INDEX TERMS Urban flood, Deep learning, Water accumulation, Real-time prediction.

show abstract

Applying gradient boosting decision trees to examine non-linear effects of the built environment on driving distance in Oslo

Cited by 186 publications

References 42 publications

Identify Risk Pattern of E-Bike Riders in China Based on Machine Learning Framework

Identify Risk Pattern of E-Bike Riders in China Based on Machine Learning Framework

When local access matters: A detailed analysis of place, neighbourhood amenities and travel choice

Real-Time Prediction of the Water Accumulation Process of Urban Stormy Accumulation Points Based on Deep Learning

Contact Info

Product

Resources

About