PDF(2422 KB)
考虑基础设施中断的暴雨灾害社会影响研究——基于公众视角的情感与行为演化分析
万欣, 丁欣宇, 张天天, 李灵芝
自然资源学报 ›› 2023, Vol. 38 ›› Issue (11) : 2919-2932.
PDF(2422 KB)
PDF(2422 KB)
考虑基础设施中断的暴雨灾害社会影响研究——基于公众视角的情感与行为演化分析
({{custom_author.role_cn}}), {{javascript:window.custom_author_cn_index++;}}Research on social impacts of rainstorm disaster considering infrastructure disruption: An analysis of sentimental and behavioral evolution from a public perspective
({{custom_author.role_en}}), {{javascript:window.custom_author_en_index++;}}利用舆情大数据探究暴雨灾害的多样化社会影响，为城市水灾害管理提供了新的研究视角。构建暴雨灾害社会影响分析框架，采集“郑州暴雨”事件中与基础设施相关的微博数据，分别运用基于词典的情感分析方法和随机森林机器学习模型，探究细粒度情感以及人与设施多种响应行为随灾情进展的演化特征。结果表明：灾前悲哀和愤怒是主导消极情绪，而灾中恐惧占主导，应及时公开灾情焦点信息，疏导高唤醒消极情绪；公众对设施中断的感知体现了级联失效关系，加强建成区积水点改造和新区交通与排水设施配套设计是关键；灾中公众对各类设施修复感知存在时滞性，应加强重点设施灾前联合防护，避免关键路段积水断行；灾前政府救助准备不足，导致公众负向行为较多，应提升基层重大风险决策与处置能力。研究结果可为防洪减灾制度及政策优化提供有益参考，有助于提升政府水灾害管理水平。
Leveraging public opinion big data to explore diverse social impacts of rainstorm and flood disasters provides a new research perspective to the management of urban water disaster emergencies. A social impact analysis framework of rainstorm disaster was established, upon which the microblog data related to critical infrastructure in the case of "Zhengzhou Rainstorm" were collected and analyzed by using lexicon-based sentiment analysis and random forest machine learning model. Thus, the evolution characteristics of fine-grained sentiments and multiple response behaviors of people and facilities were examined throughout the disaster from sentimental and behavioral dimensions. Correspondingly, the recommendations for flood control policy-making were put forward. The results show that sadness and anger were leading negative emotions before the disaster, while fear dominated during the disaster. Hence, the timely disclosure of disaster situation focus information can provide positive guidance for high arousal negative emotions. Meanwhile, the dissemination of disaster prevention knowledge is essential to enhance the public's capacity for disaster response. Public perception towards infrastructure disruption is reflected in the cascading failure relationships between diverse facilities, highlighting the criticality of water supply facilities and their sensitivity to transportation services. A key step in addressing this concern is to improve the modification of waterlogging hotspots in urban areas and the coordinated design of transportation and drainage facilities in newly-developed regions. It was also found that the restorations of facilities, such as transportation, were perceivable by the public both before and after the disaster; however there existed a time lag during the disaster. Therefore, pre-disaster joint protection of critical facilities should be reinforced to prevent waterlogging and disruption on key road sections. Moreover, the evident inadequate pre-disaster aid preparation of governments resulted in considerable negative public behaviors during this phase. Consequently, the linkage mechanism between meteorological warning and emergency response must be strengthened, while the decision-making and disposal capacity of grassroots-level governments regarding significant risks needs to be improved. This study contributes to the body of literature on disaster management and infrastructure management. The findings are beneficial to enhancing the government's capacity for water disaster management and water security by providing valuable references for optimizing flood control and disaster reduction systems and policies.
基础设施 / 社会影响 / 公众感知 / 暴雨灾害 / 水安全 {{custom_keyword}} /
infrastructure / social impacts / public perception / rainstorm disaster / water safety and security {{custom_keyword}} /
表1 基础设施分类关键词Table 1 Keywords of infrastructure classification |
| 基础设施类型 | 关键词 |
|---|---|
| 给排水设施 | 无水、停水、没水、断水、供水、有水、水力、积水、排水、污水、内涝、淹、漫过、蹚 |
| 交通设施 | 地铁、*号线、马路、路面、道路、堵塞、高架、高速、桥、隧道、公交、停运、站台、交通 |
| 电力设施 | 停电、断电、没电、无电、来电、有电、供电、用电、电线、电缆、触电、漏电、电力 |
| 通讯设施 | 断网、没网、断联、关机、有网、通讯、信号、邮政、快递、物流 |
表2 情感分类与含义Table 2 Classifications and meanings of sentiments |
| 情感类别 | 含义界定 | 示例 |
|---|---|---|
| 积极 | 对基础设施服务水平和相关部门救灾措施的满意、获救后的喜悦、公众祈福祝愿、团结一致等正向情绪 | “防汛抢险守一线乘风破浪保供电” “河南幼儿园开铲车护送孩子过积水”等 |
| 愤怒 | 对基础设施服务水平的不满、对相关部门不作为的贬责、对他人利己行为的憎恶等情绪 | “公司停电了老板不让走” “气象部门为什么没有提前通知”等 |
| 悲哀 | 对灾情发展无能为力的内疚、对救灾人员的心疼、对受灾者的担忧、对遇难者的悲伤思念、对基础设施服务的失望等情绪 | “家乡受灾我却在外地读书帮不上忙” “郑州地铁昨天有12人遇难，他们只是下班回家而已，却再也回不去了”等 |
| 恐惧 | 对突发灾害以及未知情况的慌乱、害怕等情绪 | “屋里停电了外面雨好大一直有巨响” “多户人家出现塌方多人受伤一人死亡”等 |
表4 情感分析精度检验结果Table 4 Test results of the accuracy of sentiment analysis (%) |
| 情感类别 | Precision | Recall | F-Measure |
|---|---|---|---|
| 积极 | 85.83 | 85.12 | 85.47 |
| 愤怒 | 90.76 | 87.01 | 88.85 |
| 悲哀 | 93.08 | 80.87 | 86.55 |
| 恐惧 | 76.19 | 85.41 | 80.54 |
表5 响应行为分类与含义Table 5 Classifications and meanings of the response behaviors |
| 行为类别 | 含义界定 | 具体表现 |
|---|---|---|
| 基础设施 中断行为 | 基础设施所呈现的故障和中断等状态的行为 | 给排水：停水、积水、溢洪道、管网堵塞等 交通：路面塌陷、道岔故障、路基冲毁、列车停运等 电力：供配电设施损坏、备用电源失效、区域断电等 通讯：通信设施损坏、信号中断、防汛专用通信网络缺失等 |
| 基础设施 修复行为 | 基础设施灾前接受巡检与排查、灾中和灾后的故障排除与恢复正常运行等行为 | 给排水：水位监测、积水抽排、清淤、封堵河道缺口、应急供水等 交通：路段排查、道路封闭、抢通作业、复轨、交通疏导、道路畅通等 电力：缆线加固、设备巡检、电缆抢修、临时供电、电力恢复等 通讯：缆线加固、设备巡检、基站抢修、建立应急通讯平台、调度应急通信设备、信号恢复等 |
| 政府救助行为 | 政府对受灾群众采取的转移、救援、安置等行为 | 灾前避险转移与避险知识宣传，灾中搜寻营救、医疗救治与生活保障，灾后安置与心理抚慰等 |
| 公众行为 | 公众之间利他或利己的人际互动以及反映灾害对其生产生活造成影响的行为 | 求助、互助、传谣等 |
表6 不同机器学习模型精度比较Table 6 Accuracy comparison of different machine learning models (%) |
| 机器学习模型 | Precision | Recall | F-Measure | |
|---|---|---|---|---|
| 所用模型 | 随机森林 | 82.95 | 82.88 | 82.88 |
| 基准模型 | 支持向量机 多元逻辑回归 决策树 朴素贝叶斯 K近邻 | 78.84 77.20 75.69 68.30 80.36 | 78.56 77.19 75.81 66.94 80.13 | 78.54 77.15 75.72 66.66 80.09 |
| [1] |
李禾尧, 何思源, 王国萍, 等. 国家公园灾害风险管理研究与实践及其对中国的启示. 自然资源学报, 2021, 36(4): 906-920.
自然保护地是区域灾害风险管理热点地区,国家公园是自然保护地体系的重要组成部分。国家公园管理目标的多样性决定其暴露在多种致灾因子下的承灾体多样性,使其灾害风险管理极具综合性。从国家公园管理职责、管理规划实践与科学研究热点出发,总结国家公园灾害风险管理的总体理念、共性与差异,将其置于社会—生态系统管理框架内,提出对中国国家公园体制试点区灾害风险管理的三个启示：（1）灾害风险管理是实现国家公园多重管理目标的必然需求;（2）灾害风险管理要重视维持国家公园社会—生态系统的理想状态;（3）灾害风险管理需要采用“整体”思想,在“愿景—目标—目的”的“层级式”管理目标下协调各部门,以适应性管理方式开展。
[
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [2] |
杨海峰, 翟国方. 灾害风险视角下的城市安全评估及其驱动机制分析: 以滁州市中心城区为例. 自然资源学报, 2021, 36(9): 2368-2381.
全球自然灾害以及城市化进程中的人为灾害频发,严重制约着城市的生存和安全发展。以典型的灾害多发地区滁州市中心城区为例,基于压力—状态—响应（Pressure-State-Response,PSR）概念框架,构建了各单灾种风险评价指标体系,利用耦合激励模型复合单灾种风险评估结果,定量测度了研究区的城市安全风险分布特征,并运用地理探测器对城市安全风险进行了驱动机制分析。结果表明：（1）研究区的城市安全风险等级占比从低到高分别为2.49%、8.71%、41.08%、30.47%和17.25%,以中风险占主导。在空间上,城市安全风险表现为中部高于周围边缘区域,呈现出由核心区往外逐步减弱的格局特征。（2）6个驱动因子对城市安全风险的解释力强度为人口密度（0.404）>地均GDP（0.402）>建筑承灾能力（0.095）>植被覆盖度（0.078）>路网密度（0.013）>用地类型风险（0.012）,因子交互协同作用后对结果的解释力增强。研究结果为城市安全风险评估提供了新的理论视角与研究框架,能够服务于高风险区域的灾害风险管理。
[
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [3] |
邬柯杰, 吴吉东, 叶梦琪. 社交媒体数据在自然灾害应急管理中的应用研究综述. 地理科学进展, 2020, 39(8): 1412-1422.
社交媒体是灾害应急管理的一个新兴但尚未充分利用的大数据源。然而,社交媒体数据在灾害应急管理中能发挥什么作用,目前国内的研究较少,而国外的研究亦较为分散,相关应用受限。论文对国内外利用社交媒体数据进行灾害管理相关研究进行了综述,归纳出社交媒体数据在灾害应急管理中应用的潜力、优势和问题。结果发现：① 社交媒体数据含有丰富的灾害信息,能够依靠态势感知和信息共享支持灾害预警、实时监测、损失和救助需求评估、协助快速应急响应以及预测可能的灾害风险等灾害管理工作;② 社交媒体数据的优势在于能保证灾害信息的实时性和连续性,进而提供致灾强度图、损失预测和舆情分析等结果服务于灾害应急管理;③ 地理位置信息有限、专业语料库缺乏、信息噪声处理的复杂性等原因使社交媒体数据分析结果精度受限。如何发挥社交媒体数据社会感知的优势,以弥补传统观测调查手段的不足,更加科学应对灾害应急,仍需要在社交媒体数据处理及多源数据融合分析技术方面取得突破。
[
Social media is a data source that is emerging but underutilized for disaster emergency management. There are few existing studies on the role of social media data in disaster emergency management in China, and international studies are also relatively scattered, which make relevant applications limited. This article reviewed the research on disaster management with social media data. We summarized the potential, advantages, and problems of applying social media data in disaster emergency management. The results show that: 1) Social media data are full of disaster information. They can support disaster early warning, real-time monitoring, and loss and rescue needs assessment, assist rapid emergency response, and predict possible disaster risks that relies on situational awareness and timely information. 2) The advantage of social media data is that it can ensure the timeliness and continuity of disaster information aquisition, which can provide disaster intensity maps, loss prediction, and public opinion analysis for disaster emergency management. 3) Some problems limit the accuracy of social media data analysis results, including limited geographical location information, lack of professional corpus, and complexity of information noise processing, among others. In order to make full use of the advantages of social media data to overcome the shortcomings of traditional observation and investigation methods and respond to disasters more scientifically, breakthroughs in social media data processing and multi-source data fusion analysis technology are needed. {{custom_citation.content}}
{{custom_citation.annotation}}
|
| [4] |
刘媛媛, 刘业森, 郑敬伟, 等. BP神经网络和数值模型相结合的城市内涝预测方法研究. 水利学报, 2022, 53(3): 284-295.
[
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [5] |
黄晶, 佘靖雯, 袁晓梅, 等. 基于系统动力学的城市洪涝韧性仿真研究: 以南京市为例. 长江流域资源与环境, 2020, 29(11): 2519-2529.
[
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [6] |
程晓陶, 刘昌军, 李昌志, 等. 变化环境下洪涝风险演变特征与城市韧性提升策略. 水利学报, 2022, 53(7): 757-768, 778.
[
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [7] |
黄纪心, 郭雪松. 基于应急任务驱动的灾害应对组织网络适应性机制研究: 以河南郑州“7·20”特大暴雨应对为例. 公共管理学报, 2022, 19(4): 52-64, 168-169.
[
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [8] |
张庆, 余淼. 基于两阶段混合整数规划模型的洪涝灾害应急管理研究. 运筹与管理, 2020, 29(9): 62-69.
本文以洪涝自然灾害为现实背景, 考虑多种应急物资、灾情的不确定性和应急救灾的多目标性, 集成优化灾前准备和灾后响应两阶段, 建立了一定最大救援时间下的两阶段多目标混合整数规划模型。模型的目标一是使得不同灾害情景下灾后响应阶段总物资不足惩罚和延误损失的期望最小, 目标二是使得灾前准备阶段应急物资存储点建造成本、物资存储成本及灾后响应阶段物资分配成本之和最小。该模型保证了应急救灾的及时有效以及物资的公平分配。本文设计了一种多目标遗传算法用于模型求解, 结合具体算例, 得到了模型在最大救援时间为4到9区间内任意数值下的pareto最优解, 很好地适应了决策者不同的决策需求, 并根据pareto应急方案的数目, 灾后响应阶段成本期望和两阶段总成本等模型的三个关键产出随最大救援时间的变化趋势, 得出最优的最大救援时间为5.7。
[
On the background of flood disasters, this paper considers various emergency supplies, the uncertainty of disaster events and the multiple otjectives of emergency response, combining the emergency preparation with the emergency response, and establishes a two stages bi-objective mixed integer programming model under a certain maximum rescue time. The first objective of the model is to minimize the expectation of penalty due to supply shortage and the loss owing to distribution delay in the emergency response stage under different disaster scenarios. The second objective is to minimize the total cost of the two stages. The model ensures the timely and effective response to flood disaster and the equity of supply distribution.In this paper, a multi-objective genetic algorithm is developed to solve the model and the pareto-optimal decision schemes are obtained at any maximum rescue time within the range of four to nine, which well satisfies the different needs of the decision maker. Finally, the paper suggests that the optimal maximum rescue time can be set to 5.7 from the trend in our case.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [9] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [10] |
Natural disasters and the poor management of civil engineering structures and infrastructures require timely action and new tools such as specially designed structural health monitoring platforms. This paper proposes an innovative platform based on a network of wirelessly connected, low-power, and renewable-energy-fed sensor units. The platform is a multipurpose tool for diagnostics, maintenance, and supervision, capable of simultaneously carrying out damage detection, localization, identification, and “multiclass” and “multi-material” level quantification of different types of failures. In addition, it works as a decision support tool for emergency management and post-disaster assessment, here tailored for an Italian theme park. The platform uses innovative algorithms based on the concept of the vibro-acoustic signature of the asset monitored. The vibro-acoustic signatures of the monitored assets are gathered by the microphones and accelerometers of the platform’s sensor units. Then, almost simultaneously, they are analyzed using specifically designed wavelet-based and convolutional-neural-network-based algorithms, which are able to extract crucial information about the structural and environmental conditions of both the asset and the areas of the thematic park. In addition, the platform shows escape routes during an emergency, indicating meeting points and helping people to proceed safely along a recognizable escape route to a safe place, as demonstrated by the simulations.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [11] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [12] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [13] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [14] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [15] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [16] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [17] |
冒婷婷, 孙新红, 汤育春, 等. 基于情感挖掘的基础设施社会韧性评价研究. 广西城镇建设, 2021, (2): 27-32.
[
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [18] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [19] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [20] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [21] |
中华人民共和国应急管理部. 河南郑州“7·20”特大暴雨灾害调查报告公布, https://www.mem.gov.cn/xw/bndt/202201/t20220121_407106.shtml, 2022-01-22.
Ministry of Emergency Management of the People's Republic of China. Investigation Report on "7·20" Heavy Rainstorm Disaster in Zhengzhou, https://www.mem.gov.cn/xw/bndt/202201/t20220121_407106.shtml, 2022-01-22.]
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [22] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [23] |
徐琳宏, 林鸿飞, 潘宇, 等. 情感词汇本体的构造. 情报学报, 2008, 27(2): 180-185.
[
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [24] |
安璐, 吴林. 融合主题与情感特征的突发事件微博舆情演化分析. 图书情报工作, 2017, 61(15): 120-129.
[目的/意义]微博是突发事件网络舆情传播的重要媒介。面向突发事件的微博主题和情感挖掘对掌握突发事件的网络舆情、识别与预测潜在问题及风险等方面具有重要的实践意义。尝试提出一种融合主题与情感特征的突发事件微博舆情演化分析方法。[方法/过程]以寨卡事件为例，通过划分微博舆情演化的生命周期，基于word2vec技术分别提取该事件生命周期各阶段的微博主题，采用基于词典的情感分析方法，引入情感词、表情符号等多情绪源，对不同主题下的评论情感做细粒度划分，并计算情感强度，最终实现微博主题与情感的协同分析。[结果/结论]所提出的微博舆情演化分析方法能够揭示面向特定事件的微博在突发事件生命周期各阶段的主题特征、情感类型与强度，剖析网络舆情主题与情感特征的协同演化规律。
[
<strong>[Purpose/significance]</strong> Microblog is an important medium of communicating public opinions toward public emergencies. Excavating topics and sentiment of microblogs on public emergencies has important practical significance to grasp the online public opinions, and to identify and predict the potential problems and risks during public emergencies. In this study, we propose an approach to analyze the topical and emotional evolution of microblog public opinions on public emergencies. <strong>[Method/process]</strong> The Zika outbreak is taken as an example and the life cycle of related microblog public opinions is divided into several phases. Topics are extracted from microblogs by the word2vec technique. The sentiment analysis has been conducted based on dictionaries containing sentiment words and emoticons to categorize the sentiment of microblogs at a fine-grained level. The emotional intensity of microblogs for each topic is also calculated to achieve synergetic analysis of topics and sentiment of microblogs. <strong>[Result/conclusion]</strong> The proposed method can reveal the topical and emotional features and emotional intensity of microblogs on specific public emergencies and illustrate the synergetic evolution patterns of topics and sentiment of online public opinions.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [25] |
林筱妍, 吴升. 基于语义规则和词向量的台风灾害网络情感分析方法. 地球信息科学学报, 2022, 24(1): 114-126.
灾害期间的舆情引导有助于维护社会稳定。社交媒体是舆论传播的重要渠道,通过微博评论了解用户的网络情感及关注的话题,可以帮助相关舆情监测部门掌握公众的关注热点,从而选择适当的干预节点来应对网络舆情,并对公众情绪进行疏导,这对于应急管理具有现实意义。现有的研究大多是利用有监督的机器学习方法进行情感分类,这需要人工进行语料的标注,工作量大。本文根据微博评论文本的特点,综合考虑情感词以及表情符号等多重情感源,构建了台风灾害领域情感词典。在此基础上,提出了一种基于情感词语义规则的情感倾向计算方法,以及基于词向量的话题聚类方法。首先,采集了近年5次台风灾害期间共计40多万条微博评论文本,基于大连理工情感词汇本体库进行扩展构建了台风灾害领域情感词典,结合PMI法构建表情符号词典,根据语义规则确定情感倾向,并使用3500条评论文本验证了该方法的有效性。然后,本文基于词向量、TF-IDF与K-means的聚类方法探索灾害期间热点话题。最后,以2020年4号台风“黑格比”为例,基于台风期间的5万余条微博评论文本进行了舆情情感分析,并识别出6类与台风相关的话题。通过时空分析发现,随着时间的推移,微博评论文本的数量发生一定变化,评论数量多的地区大都集中在沿海地区和经济水平高的地区,台风登陆当天浙江省的恐惧情感达到最高。结果表明,基于语义规则和词向量的台风灾害网络情感分析方法,能在类似灾害事件发生时为政府部门掌握和引导网络舆情提供辅助。
[
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [26] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [27] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [28] |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [29] |
许乙青, 喻丁一, 冉静. 基于流域协同的国土空间雨洪安全格局构建方法. 自然资源学报, 2021, 36(9): 2335-2349.
近年来愈发频繁的洪涝灾害暴露出城市建设与雨洪安全之间的矛盾已不容忽视,以防洪基础设施为核心的防灾规划和以各类景观生态基础设施为核心的海绵城市规划,需要内涵和方法的转变。以“多规协同”为理论基础,提出在国土空间规划背景下,雨洪安全格局构建的核心内涵应该是“协同关系”的构建,即不同规划要素之间不仅在空间布局上协调无冲突,更重要的是能够导向同一的目标与结果。基于该内涵,提出了国土空间雨洪安全格局构建方法的理论框架,应以雨洪安全要素的识别为前提,通过强制性约束措施和引导性联动措施构建雨洪安全要素与国土空间结构、资源用途和支撑系统的协同关系,最终实现不同空间尺度上雨洪安全目标的落实与格局的形成。为更好地诠释该理论框架,以江西省万载县城关镇镇域国土空间规划为案例,探索并检验该理论框架在实践中的可操作性及应用价值,以期为落实国土空间雨洪安全格局的目标提供新的思路和理论依据。
[
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| [30] |
周利敏, 谭妙萍. 中国灾害治理: 组织, 制度与过程研究综述. 理论探讨, 2021, (6): 138-146.
[
{{custom_citation.content}}
{{custom_citation.annotation}}
|
| {{custom_ref.label}} |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
PDF(2422 KB)
图1 研究框架
表1 基础设施分类关键词图2 基础设施相关微博发文量图3 灾害各阶段基础设施相关微博词云表2 情感分类与含义表3 否定词修饰对情感类别的转化规则表4 情感分析精度检验结果表5 响应行为分类与含义表6 不同机器学习模型精度比较图4 四类情感占比及随时间演化图5 消极情感演化图6 基础设施中断行为演化图7 基础设施修复行为演化图8 政府救助行为与公众行为演化图9 公众行为情感演化/
| 〈 |
|
〉 |
AI Summary
