High speed rail passenger segmentation and ticketing channel preference

A multi-channel strategy for the purposes of marketing and ticketing has been widely employed in various fields, including the transportation industry, yet few transportation studies have investigated ticket channel-related issues. This study thus adopts Taiwan High Speed Rail (THSR) as a case study to identify passengers’ perceptions regarding key factors that affect the channel preference through which they receive their services, particularly across a four-stage ticket purchasing process, including information enquiry, booking, payment, and ticket pick-up. To investigate these key factors, we base the framework of our research model upon customers’ perceived value perspective, adopting a multinomial logit model to examine the influence of channel attributes on HSR passengers’ preferences. The relationship among the decisions made by these passengers at each ticket purchasing stage is also examined. The findings of our study demonstrate that perceived risk, perceived benefit, and perceived ease of use are critical factors influencing passengers’ channel preference throughout the ticket-purchasing process. Perceptual differences are proven to exist due to various demographic factors and trip characteristics. The conclusions of this study have implications for THSR management to design appropriate ticketing channel services for certain types of passengers and can also be generally applied to multiple distribution channels in conventional railway system and intercity bus services.     

城市快速路区间旅行时间短期预测新算法

城市道路交通状况的预测,是实现未来路况查询、车辆动态导航等智能交通系统技术的关键。该文在分析浮动车数据的时间相关性的基础上,研究城市快速路的区间旅行时间短期预测算法。首先,采用统计方法和K-NN分类法相结合的新方法对缺失数据进行填充,并利用小波变换对每天的数据进行消噪处理;其次,在分别利用时间序列模型和人工神经网络模型对城市快速路区间旅行时间进行短期预测的基础上,通过模型组合获得预测值;最后,结合北京市区二环的一段快速路区间旅行时间的历史数据和实时数据,对该文所提出的快速路区间旅行时间短期预测算法进行了评价。结果显示,该算法的预测结果的平均绝对误差百分比控制在10.43%以内,具有良好的精度。     

Online sequential extreme learning machine based multilayer perception with output self feedback for time series prediction

This study presents a time series prediction model with output self feedback which is implemented based on online sequential extreme learning machine. The output variables derived from multilayer perception can feedback to the network input layer to create a temporal relation between the current node inputs and the lagged node outputs while overcoming the limitation of memory which is a vital part for any time-series prediction application. The model can overcome the static prediction problem with most time series prediction models and can effectively cope with the dynamic properties of time series data. A linear and a nonlinear forecasting algorithms based on online extreme learning machine are proposed to implement the output feedback forecasting model. They are both recursive estimator and have two distinct phases: Predict and Update. The proposed model was tested against different kinds of time series data and the results indicate that the model outperforms the original static model without feedback.     

基于时间序列法的2014年乘用车月销量预测

对我国乘用车整体及其细分车型月度销量进行合理预测,对于企业合理安排生产,优化市场规划有着重要的意义。文章以2005年1月-2013年10月共计106个月的历史数据为依据,运用时间序列方法和StatTools工具对我国2014年度乘用车及细分车型月度销量进行了预测。预测结果准确度高,具有较高的参考价值。采用时间序列法进行预测,仅以客观数据为依据,结果更加准确合理,为整车企业安排生产提供了参考。     

基于小波神经网络的道路交通流量实时预测模型研究

实时交通流量预测是智能交通系统的核心内容,智能交通系统中多个子系统的功能实现都以其为基础。交通流具有高度非线性和不确定等特征,且与时间高度相关,可以看成是时间序列预测问题。根据交通流的这些特点,提出基于小波神经网络的道路交通流量实时预测模型,并以某条道路为例,通过Matlab编程实现模拟仿真。仿真结果表明,小波神经网络能够比较精确、快速地对实时交通流量进行预测,网络预测值接近期望值。     

基于小波变换的海浪方向谱估计方法研究*

利用数值模拟方法分析了基于小波变换的方向谱估计分析方法的适用性,测波阵列类型、波浪入射方向、方向 集中度以及采用长度对小波方向谱估计分析方法分析结果的影响。结果表明：小波方向谱的方法适用条件较广泛,实用 性更强。     

The interurban DRAG-Spain model: The main factors of influence on road accidents in Spain

This paper presents the results of applying DRAG methodology to the identification of the main factors of influence on the number of injury and fatal accidents occurring on Spain’s interurban network. Nineteen independent variables have been included in the model grouped together under ten categories: exposure, infrastructure, weather, drivers, economic variables, vehicle stock, surveillance, speed and legislative measures. Highly interesting conclusions can be reached from the results on the basis of the different effects of a single variable on each of the accident types according to severity. The greatest influence revealed by the results is exposure, which together with inexperienced drivers, speed and an ageing vehicle stock, have a negative effect, while the increased surveillance on roads, the improvement in the technological features of vehicles and the proportion of high capacity networks have a positive effect, since the results obtained show a significant drop in accidents.     

Segmentation of paratransit users based on service quality and travel behaviour in Bandung,Indonesia

Using a data-set collected among paratransit users in Bandung, Indonesia, this paper explores the impacts of paratransit users' negative experiences and dissatisfactions with their paratransit usage pattern. Segmentation and ordered probit analyses are used to examine the impacts of users' opinions on service quality on their trip-making behaviour. The results indicate that users are divided into six segments – namely, the unlucky, the young user, the experienced, the adapter, the infrequent user and the captive. The results further indicate that paratransit users, especially women, perceive negative experiences related to on-time performance (departure and arrival time) and security issues related to vehicles. Despite these dissatisfactions and negative experiences, they are still likely to use paratransit as part of their daily life. Two paratransit market segments (the adapter and the captive) are also likely to result in more trips using paratransit. The study offers several strategies and recommendations that can improve the current paratransit system so that it can better serve local needs.     

基于萤火虫优化的BP神经网络算法研究

为了改进神经网络结构和参数的设置方法,在萤火虫算法和BP神经网络的基础上,提出了一种萤火虫算法优化BP神经网络的算法.该算法利用萤火虫算法得到更优的网络初始权值和阈值,弥补BP神经网络连接权值和阈值选择上的缺陷.将该算法应用到Duffing系统产生的混沌时间序列进行算法的有效性验证,并与BP神经网络进行比较,仿真结果表明该算法具有更高的预测准确性,从而证明该算法在该预测领城的可行性和有效性.     

舰艇特殊进水过程的流量差值迭代算法研究

目前,在进行静水情况下的破损舰艇进水过程中的姿态计算时,通常使用流量差值迭代算法。然而,该算法的使用有一定的局限性。对于存在进满水舱室的特殊进水过程,破口流线的伯努利方程将发生改变,这导致了不能使用目前的迭代公式进行计算。因此,本文以该特殊的进水过程为对象,在对流线伯努利方程修正的基础上,利用隐函数求导模型和克莱姆法则,对此时的流量差值迭代算法进行了建模,并对该算法进行了船模实验验证。最后,针对某船的进水过程进行了仿真计算,横倾角时域变化曲线上的不连续点直观地表明了流量差值迭代算法的改变对进水过程仿真计算的影响。