Applicability Analysis and Ensemble Application of BERT with TF-IDF, TextRank, MMR, and LDA for Topic Classification Based on Flood-Related VGI DOI Creative Commons
Wenying Du, Chang Ge, Shuang Yao

et al.

ISPRS International Journal of Geo-Information, Journal Year: 2023, Volume and Issue: 12(6), P. 240 - 240

Published: June 9, 2023

Volunteered geographic information (VGI) plays an increasingly crucial role in flash floods. However, topic classification and spatiotemporal analysis are complicated by the various expressions lengths of social media textual data. This paper conducted applicability on bidirectional encoder representation from transformers (BERT) four traditional methods, TextRank, term frequency–inverse document frequency (TF-IDF), maximal marginal relevance (MMR), linear discriminant (LDA), results show that for user type, BERT performs best Government Affairs Microblog, whereas LDA-BERT We Media Microblog. As text length, TF-IDF-BERT works better texts with a length <70 >140 words, 70–140 words. For evolution pattern, study suggests Henan rainstorm, topics follow general pattern “situation-tips-rescue”. Moreover, this detected hotspot “Metro Line 5” related to rainstorm discovered topical focus spatially shifts Zhengzhou, first Xinxiang, then Hebi, showing remarkable tendency south north, which was same as report issued authorities. integrated multi-methods improve overall accuracy Sina microblogs, facilitating flooding.

Language: Английский

Hydraulic modelling of inland urban flooding: Recent advances DOI Creative Commons
Emmanuel Mignot, Benjamin Dewals

Journal of Hydrology, Journal Year: 2022, Volume and Issue: 609, P. 127763 - 127763

Published: March 25, 2022

Language: Английский

Citations

72
Knowledge and topology: A two layer spatially dependent graph neural networks to identify urban functions with time-series street view image DOI
Yan Zhang, Pengyuan Liu, Filip Biljecki

et al.

ISPRS Journal of Photogrammetry and Remote Sensing, Journal Year: 2023, Volume and Issue: 198, P. 153 - 168

Published: March 16, 2023

Language: Английский

Citations

45
Mapping the landscape and roadmap of geospatial artificial intelligence (GeoAI) in quantitative human geography: An extensive systematic review DOI Creative Commons
Siqin Wang, Xiao Huang, Pengyuan Liu

et al.

International Journal of Applied Earth Observation and Geoinformation, Journal Year: 2024, Volume and Issue: 128, P. 103734 - 103734

Published: March 11, 2024

This paper brings a comprehensive systematic review of the application geospatial artificial intelligence (GeoAI) in quantitative human geography studies, including subdomains cultural, economic, political, historical, urban, population, social, health, rural, regional, tourism, behavioural, environmental and transport geography. In this extensive review, we obtain 14,537 papers from Web Science relevant fields select 1516 that identify as studies using GeoAI via scanning conducted by several research groups around world. We outline applications systematically summarising number publications over years, empirical across countries, categories data sources used applications, their modelling tasks different subdomains. find out existing have limited capacity to monitor complex behaviour examine non-linear relationship between its potential drivers—such limits can be overcome models with handle complexity. elaborate on current progress status within each subdomain geography, point issues challenges, well propose directions opportunities for future context sustainable open science, generative AI, quantum revolution.

Language: Английский

Citations

37
Urban flood susceptibility mapping based on social media data in Chengdu city, China DOI Creative Commons
Yao Li, Frank Osei, Tangao Hu

et al.

Sustainable Cities and Society, Journal Year: 2022, Volume and Issue: 88, P. 104307 - 104307

Published: Nov. 17, 2022

Increase in urban flood hazards has become a major threat to cities, causing considerable losses of life and the economy. To improve pre-disaster strategies mitigate potential losses, it is important make susceptibility assessments carry out spatiotemporal analyses. In this study, we used standard deviation ellipse (SDE) analyze spatial pattern floods find area interest (AOI) based upon related social media data that were collected Chengdu city, China. We as response variable selected 10 flood-influencing factors independent variables. estimated model using Naïve Bayes (NB) method. The results show events are concentrated northeast-central part especially around city center. Results checked by Receiver Operating Characteristic (ROC) curve, showing under curve (AUC) was equal 0.8299. This validation result confirmed can predict with satisfactory accuracy. map center provides realistic reference for monitoring early warning.

Language: Английский

Citations

65
Navigating Post-COVID-19 Social–Spatial Inequity: Unravelling the Nexus between Community Conditions, Social Perception, and Spatial Differentiation DOI Creative Commons

Minjun Zhao,

Ning Liu, Jinliu Chen

et al.

Land, Journal Year: 2024, Volume and Issue: 13(4), P. 563 - 563

Published: April 22, 2024

The 2023 SDGs report underscores the prolonged disruption of COVID-19 on community living spaces, infrastructure, education, and income equality, exacerbating social spatial inequality. Against backdrop dual impact significant events emergence digital technologies, a coherent research trajectory is essential for characterizing social–spatial equity understanding its influential factors within urban planning discipline. While prior emphasized dimensions mitigated differentiation to ensure equity, complexity these interconnections necessitates more comprehensive approach. This study adopts holistic perspective, focusing “social–spatial” dynamics, utilizing perception (sentiment maps) (housing prices index) pre- post-pandemic elucidate interconnected interactive nature uneven development at scale. It employs multi-dimensional methodological framework integrating morphology analysis housing conditions, GIS amenities, sentiment semantic public opinion, multiscale geographically weighted regression (MGWR) correlation factors. Using Suzhou, China, as pilot study, this demonstrates how integrated methods complement each other, exploring conditions resource distribution collectively bolster resilience, thereby maintaining amidst pandemic disruptions. findings reveal that exacerbates stratification differentiation. proximity well-maintained ecological environments, such parks or scenic landmarks, generally exhibits consistency positive effects measurement. Simultaneously, various elements influencing show geographic heterogeneity, particularly in areas farther from central regions Xiangcheng Wujiang districts. uncovers bilateral mechanism between differentiation, aiming delve into interdependent relationship built environmental Furthermore, it aspires provide meaningful references recommendations regeneration policy formulation era sustain equity.

Language: Английский

Citations

16
RB-TRNet: a regularity-guided and boundary-aware architecture for toponym recognition from Chinese text DOI Creative Commons
Haigang Sui, J. J. Wang, Xining Zhang

et al.

Geo-spatial Information Science, Journal Year: 2025, Volume and Issue: unknown, P. 1 - 15

Published: Jan. 17, 2025

Extracting geographic information from texts contributes to both science research and various practical applications, but extracting fine-grained complex location descriptions Chinese text is still challenging, due flexible word construction lack of clear boundaries in place names. In this paper, we propose a regularity-guided boundary-aware architecture for toponym recognition (RB-TRNet), achieving name by learning the internal compositional patterns constructions automatically perceiving types entities. First, RoBERTa used represent input containing Then, two BiLSTM layers are fed with representation sequences, one processed sequence entering module obtain composition entities other regularity-discriminant soften an excessive reliance on contextual recognizing Additionally, orthogonal space established after network facilitate different rule features modules. Finally, joint optimization training three modules, regularity perception predict To validate results, new (CCPNT) dataset recognition. The CCPNT dataset, along public datasets, were performance evaluation, compared eight baseline models, RB-TRNet exhibited state-of-the-art

Language: Английский

Citations

1
Analysis of Urban Flooding Driving Factors Based on Water Tracer Method and Optimal Parameters-Based Geographical Detector DOI Creative Commons
Kui Xu, Yong Tian, Lingling Bin

et al.

International Journal of Disaster Risk Science, Journal Year: 2025, Volume and Issue: unknown

Published: April 10, 2025

Abstract Urban flooding is caused by multiple factors, which seriously restricts the sustainable development of society. Understanding driving factors urban pivotal to alleviating flood disasters. Although effects various on have been extensively evaluated, few studies consider both interregional connection and interactions between factors. In this study, were analyzed based water tracer method optimal parameters-based geographical detector (OPGD). An simulation model coupled with was constructed simulate flooding. Furthermore, volume results. Subsequently, force them quantified using OPGD model. Taking Haidian Island in Hainan Province, China as an example, results show that sub-catchment H6 region experiencing most severe H9 contributes overall study area. The subsequent effect analysis elevation factor maximum single-factor (0.772) ∩ percentage building area pair two-factor (0.968). addition, bivariable or nonlinear enhancement effects. two strengthen influence each This understanding cause provides a reference for risk mitigation.

Language: Английский

Citations

1
Urban Flood-Related Remote Sensing: Research Trends, Gaps and Opportunities DOI Creative Commons
Wei Zhu, Zhe Cao, Pingping Luo

et al.

Remote Sensing, Journal Year: 2022, Volume and Issue: 14(21), P. 5505 - 5505

Published: Nov. 1, 2022

As a result of urbanization and climate change, urban areas are increasingly vulnerable to flooding, which can have devastating effects on the loss life property. Remote sensing technology provide practical help for flood disaster management. This research presents review flood-related remote identify trends gaps, reveal new opportunities. Based Preferred Reporting Items Systematic Reviews Meta-Analyses (PRISMA), systematic literature search resulted in 347 documents classified as geography, management application, data utilization. The main results include 1. most studies located high-income countries territories inland areas; 2. observing environment was more popular than building; 3. often applied activities were vulnerability assessment risk modeling (mitigation) rapid damage (response); 4. DEM is simulate floods software inputs. We suggest that future directions coastal study non-high-income countries/territories populations; understudied activities, need observe buildings standardization will facilitate integration with international standard methods assessing floods.

Language: Английский

Citations

35
City2vec: Urban knowledge discovery based on population mobile network DOI
Yan Zhang, Xiang Zheng, Marco Helbich

et al.

Sustainable Cities and Society, Journal Year: 2022, Volume and Issue: 85, P. 104000 - 104000

Published: July 13, 2022

Language: Английский

Citations

31
Rapid urban flood risk mapping for data-scarce environments using social sensing and region-stable deep neural network DOI
Lin Lin, Chaoqing Tang, Qiuhua Liang

et al.

Journal of Hydrology, Journal Year: 2022, Volume and Issue: 617, P. 128758 - 128758

Published: Nov. 24, 2022

Language: Английский

Citations

30