Using machine learning and big data for efficient forecasting of hotel booking cancellations

Cancellations are a key aspect of hotel revenue management because of their impact on room reservation systems. In fact, very little is known about the reasons that lead customers to cancel, or how it can be avoided. The aim of this paper is to propose a means of enabling the forecasting of hotel booking cancellations using only 13 independent variables, a reduced number in comparison with related research in the area, which in addition coincide with those that are most often requested by customers when they place a reservation. For this matter, machine-learning techniques, among other artificial neural networks optimised with genetic algorithms were applied achieving a cancellation rate of up to 98%. The proposed methodology allows us not only to know about cancellation rates, but also to identify which customer is likely to cancel. This approach would mean organisations could strengthen their action protocols regarding tourist arrivals.     

云南省团队游空间结构及其聚散效应研究

在旅游业高质量快速发展的背景下,全面厘清与掌握省域尺度范围各旅游节点在网络中的角色和地位,并解析其网络结构效应具有重大意义。针对现有研究因旅游流数据采集限制而较少对团队游空间结构研究的不足,基于云南省旅游包车客运信息,运用社会网络结合空间分析的方法,提取了云南省团队游网络空间结构特征,并对其聚散效应进行分析。结果表明:云南省团队游整体网络密度较小且淡旺季表现出相似特征,网络对各节点的旅游联系与合作影响尚未显现,暴露出云南省团队游旅游市场发展的不均衡;团队游网络外向中心势显著高于内向中心势,网络空间集聚趋势明显;通常具有较强外向中心性的旅游节点同时也具有较强内向中心性,表现出“集中—发散”的聚散规律;旅游节点的聚散效应表现为单个节点以圈层状向外递减形式发生,其扩散路径主要集中于文山—蒙自、景洪—思茅方向。整体上,云南省区域旅游业发展差异明显,应重点加强滇西南与滇西旅游区、滇中与滇东北旅游区的旅游资源开发与合作。     

Heterogeneity and netting efficiency under central clearing: A stochastic network analysis

This paper examines the effect of heterogeneity in clearing members’ exposure management practices under central clearing. Our network model specifies the dynamics of prenetted interbank exposures to shape interdependent exposure distributions beyond normality. Employing over-the-counter derivatives market data from the U.S. Office of the Comptroller of the Currency, our simulation results indicate that heterogeneity in bank-to-bank exposure dynamics is systemically desirable, while the entire system benefits more from the central clearing in more homogeneous environments. Policymakers should incentivize individual clearing members to enhance resiliency and stability in counterparty exposure management to maximize netting efficiency under central clearing.     

Tariffs and formation of free trade agreements networks

This paper examines the formation of bilateral free trade agreements (FTAs) on the basis of country heterogeneity in the tariff level. We demonstrate that a country's unilateral incentive to form an FTA depends on the relative magnitudes of the (negative) market concession effect and the (positive) market expansion effect, both of which are determined by the tariff levels of the two FTA partner countries. Global welfare is maximised when all country pairs form FTAs. Two countries in equilibrium are more likely to form an FTA when their tariff gap is smaller or when their tariff levels are neither very high nor very low. This finding is robust to several extensions of the model. Our preliminary empirical analysis provides some evidence for the finding.     

基于机器学习的斜坡堤越浪量预测方法研究

针对斜坡堤越浪量预测方法,分别建立集成神经网络(ensemble neural network,ENN)、随机森林(random for-eset,RF)和支持向量回归机(suppport vector regression,SVR)3种机器学习模型对斜坡堤越浪量进行预测,并利用决定系数R2和均方根误差RMSE来评估模型性能.最后,对3种模型的性能进行分析.结果显示,集成神经网络模型的决定系数R2和均方根误差RM S E分别约为0.96和0.0018,随机森林模型的决定系数R2和均方根误差RMSE分别约为0.97和0.0014,支持向量回归机模型的决定系数R2和均方根误差RMSE分别约为0.94和0.002.对比发现,3种模型的决定系数都达到0.9以上,都具有较高的预测精度,随机森林相比其他两个模型精度更高.     

基于ANP-PP-SPA的区域洪灾风险评价模型研究

针对现有洪灾风险评价模型的缺陷与不足,从致灾因子、孕灾因子、承灾因子和减灾因子4个方面,通过对14项评价指标的分析与计算,构建洪灾风险评价体系。洪灾风险评价体系以网络层次分析法(ANP)求解主观权重,投影寻踪法(PP)求解客观权重,主客观综合权重与集对分析理论(SPA)耦合,构建基于ANP-PP-SPA的洪灾风险评价模型。以广东省英德市为例,验证洪灾风险评价模型的适用性。计算结果表明,英德市2016年的洪灾风险属于中等级别,符合英德市2016年的实际情况。洪灾风险评价模型不仅综合考虑了评价指标间的相互关系,还较好地体现了洪灾风险的模糊性和随机性,能够为区域洪灾风险决策和洪水管理提供科学依据。     

Parking futures: An international review of trends and speculation

The explosion of low-cost, on-demand taxi services and the anticipation of an autonomous vehicle future has made transportation the center of debate and discussion for the first time since the massive expansion of the US highway system in the 1950s. Yet the realm of parking boasts innovations and developments far beyond the high-profile issues of TNCs and AVs. Rather, innovation in parking is happening in many cases quietly on a wide variety of fronts, including technology, public policy, and design. This paper serves an overview of emerging trends in parking, primarily within the US context. We identify and outline five developments and the pertinent technologies helping to catalyze change: unbundling parking costs, reducing parking minimums, pricing and allocating curb space dynamically, designing hybrid parking structures, and preparing for the autonomous era and “mobility as a service.” This paper presents these trends with illustrative examples highlighting current practices, governance challenges, and possible future scenarios.     

Beyond the community in participatory forest management: A governance network perspective

Governance of the environment and natural resources involves interests of multiple stakeholders at different scales. In community-based forest management, organisations outside of communities play important roles in achieving multiple social and ecological objectives. How and when these organisations play a role in the community-based forest management process remains a key question. We applied social network analysis to a case study in Indonesian Borneo to better understand the evolution of interactions between organisational actors, and with communities. NGOs featured most prominently in initiating the permit process, implementing management, and providing other support activities, while also being well-connected to donors and government actors. The network configurations indicated significant cooperation among organisations when initiating the community forest process, while bridging between village and organisational levels characterised all stages of the community forest process. While community-based forest management often evokes images of grassroots efforts and broad local capacity to manage forests, reality shows a more dynamic and heterogeneous picture and broader involvement of different actor types and motivations in Indonesia. These findings can be applied to other countries implementing and expanding their decentralised forest policies.     

深度学习技术在海关风险甄别中的应用研究

随着互联网的高速发展和人工智能时代的到来,越来越多从前必须由人脑完成的工作能够利用计算机技术来完成,而深度学习的出现更解决了传统机器学习算法在计算机视觉领域、自然语言处理领域表现不佳的问题,使机器也能够拥有准确感知图像和语音的能力。人脸识别是深度学习网络最常见的应用场景之一,具有自然、直接、方便的特点,且不需要检测对象配合,因此非常适合用于公共安全领域的风险检测。研究充分结合海关实际需求,搭建基于深度学习技术的人脸识别模型,提供对通关旅客进行实时风险甄别的解决方案,以及海关通关风险防控场景的理论参考,为后续深度学习技术在海关业务的研究提供支撑。     

Systemic financial risk early warning of financial market in China using Attention-LSTM model

We propose an Attention-LSTM neural network model to study the systemic risk early warning of China. Based on text mining, the network public opinion index is constructed and used as a training set to be incorporated into the early warning model to test the early warning effect. The results show that: (i) the network public opinion is the non-linear Granger causality of systemic risk. (ii) The Attention-LSTM neural network has strong generalization ability. Early warning effects have been significantly improved. (iii) Compared with the BP neural network model, the SVR model and the ARIMA model, the LSTM neural network early warning model has a higher accuracy rate, and its average prediction accuracy for systemic risk indicators has been improved over short, medium and long terms. When the attention mechanism is included in the LSTM, the Attention-LSTM neural network model is even more accurate in all the cases.