A machine-learning method for improving crash injury severity analysis: a case study of work zone crashes in Cairo,Egypt

Abstract

The quality of vehicular collision data is crucial for studying the relationship between injury severity and collision factors. Misclassified injury severity data in the crash dataset, however, may cause inaccurate parameter estimates and consequently lead to biased conclusions and poorly designed countermeasures. This is particularly true for imbalanced data where the number of samples in one class far outnumber the other. To improve the classification performance of the injury severity, the paper presents a robust noise filtering technique to deal with the mislabels in the imbalanced crash dataset using the advanced machine learning algorithms. We examine the state-of-the-art filtering algorithms, including Iterative Noise Filtering based on the Fusion of Classifiers (INFFC), Iterative Partitioning Filter (IPF), and Saturation Filter (SatF). In the case study of Cairo (Egypt), the empirical results show that: (1) the mislabels in crash data significantly influence the injury severity predictions, and (2) the proposed M-IPF filter outperforms its counterparts in terms of the effectiveness and efficiency in eliminating the mislabels in crash data. The test results demonstrate the efficacy of the M-IPF in handling the data noise and mitigating the impacts thereof.     

一种基于预测值量测转换的卡尔曼滤波跟踪算法

在雷达目标跟踪中，系统量测信息通常在球坐标系下获得。为了采用经典卡尔曼滤波算法实现有效目标跟踪，通常采用量测转换方法将非线性量测信息转换到直角坐标系中。针对传统量测转换方法基于量测值计算转换误差统计特性而导致的估计结果有偏问题，提出了一种基于预测值的量测转换方法，并将其与卡尔曼滤波算法相结合，获得了一种基于预测值量测转换的卡尔曼滤波跟踪算法。仿真结果表明，与现有的基于量测转换的卡尔曼滤波算法相比，该算法能在不提高运算量的情况下有效改善目标跟踪效果，跟踪精度提升约20%。     

一种基于Edline线特征的车道线识别算法

为了提升车道线检测的准确性和实时性,改良车道偏离预警系统的性能,提出了一种新的车道线识别算法。首先应用投影法对采集到的图像设立感兴趣区域,以此来减少图像中存在的干扰信息;其次应用一种改进后的自适应高斯滤波算法对采集所得图像进行平滑处理,减少图像中不必要的细节;最后采用边缘绘制算法进行边缘检测,在此基础上,提出一种线段检测算法——Edline算法提取边缘线,对检测到的直线段进行筛选和聚类。利用引用计数法对车道线进行跟踪和预测。结果表明,新算法的平均处理时间为17.1 ms,准确率为96.19%,将其应用在车道偏离预警系统中可以有效地提高预警效率,提升预警的准确性和响应速度。研究结果丰富了车道线识别理论,可为车道偏离预警系统的应用及基础研究提供参考。     

一种多传感器空情信息融合算法

针对防空指挥系统快速反应决策能力要求,给出了可行的系统信息融合处理软件组成结构及功能描述,提出了一种适用于快速空情信息融合的通用信息融合处理方法和模型算法。建模过程中,提出的信息融合处理方法和模型算法将防空指挥系统信息融合的复杂过程条理化、公式化和数字化。仿真结果证明,该方法是正确和有效的。目前,该方法已应用到实际工程中。     

The effects of choice set size and information filtering mechanisms on online hotel booking

Online booking is one of the most popular ways of making reservations for hotel guests. Thus, hoteliers are paying increasing attention to hotel website information presentation and design. The purpose of this study, then, is to examine the joint influence of choice set size and information filtering mechanisms on consumers’ decision confidence towards online hotel booking. Choice set size was operationalized through 3-, 9-, and 30-hotel room choice sets. Through experimental design, this study shows that the presence of an information filtering mechanism reduces consumers’ perceptions of choice overload with a large number of choices (30 choices), whereas its impact is attenuated with smaller choice sets (3 and 9 choices). In addition, choice overload mediates the impact of choice set size on decision confidence. Theoretical contribution and managerial implications are also discussed.     

PRICE SETTING OF MARKET MAKERS: A FILTERING PROBLEM WITH ENDOGENOUS FILTRATION

This paper studies the price‐setting problem of market makers under risk neutrality and perfect competition in continuous time. The classic approach of Glosten–Milgrom is followed. Bid and ask prices are defined as conditional expectations of a true value of the asset given the market makers' partial information that includes the customers' trading decisions. The true value is modeled as a Markov process that can be observed by the customers with some noise at Poisson times. A mathematically rigorous analysis of the price‐setting problem is carried out, solving a filtering problem with endogenous filtration that depends on the bid and ask price processes quoted by the market maker. The existence and uniqueness of the bid and ask price processes is shown under some conditions.     

Nonlinear causality testing with stepwise multivariate filtering: Evidence from stock and currency markets

We examine the spillovers of the US subprime crisis to Asian and European economies and in particular to what extent currency and stock markets have been affected by the crisis. Linear and nonlinear dependencies are detected after pairwise and system-wise causality analysis. A new stepwise multivariate filtering approach is implemented after controlling for conditional heteroskedasticity in the raw data and in VAR/VECM residuals using multivariate GARCH models. Significant nonlinear causal linkages persisted even after the application of GARCH-BEKK, CCC-GARCH and DCC-GARCH modelling. This indicates that volatility effects might partly induce nonlinear causality. Perhaps new short-term asset-pricing models could be developed to explain this stylized fact. These results might also have important implications for hedging, trading strategies and financial market regulation.     

动液面谱减法滤波

动液面滤波是油井检测动液面的关键技术之一。文章利用谱减法处理油井动液面信号,消除背景噪声的影响,并改进谱减法应用到动液面滤波,取得较好的滤波效果,从而提高对液面反射波位置的辨识能力。     

The European refugee crisis and the natural rate of output

The European Commission follows a harmonized approach for calculating structural (potential) output for EU member states that takes into account labour as an important ingredient. This article shows how the recent huge migrants’ inflow to Europe affects trend output. Due to the fact that the immigrants immediately increase the working population but effectively do not enter the labour market, we illustrate that the potential output is potentially upward biased without any corrections. Taking Germany as an example, we find that the average medium-term potential growth rate is lower if the migration flow is modelled adequately compared to results based on the unadjusted European Commission procedure.     

Personalized multi-period tour recommendations

During a trip planning, tourists gather information from different sources, select and rank the places to visit according to their personal interests, and try to devise daily tours among them. This paper addresses the complex selection and touring problem and proposes a “filter-first, tour-second” framework for generating personalized tour recommendations for tourists based on information from social media and other online data sources. Collaborative filtering is applied to identify a subset of optional points of interest that maximize the potential satisfaction, while there are some preselected mandatory points that the tourists must visit. Next, the underlying orienteering problem is solved via an Iterated Tabu Search algorithm. The goal is to generate tours that contain all mandatory points and maximize the total score collected from the optional points visited daily, taking into account different day availabilities and opening hours, limitations on the tour lengths, budgets and other restrictions. Computational experiments on benchmark datasets indicate that the proposed touring algorithm is very competitive. Furthermore, the proposed framework has been evaluated on data collected from Foursquare. The results show the practical utility and the temporal efficacy of the recommended tours.