Multimodal Causal Inference Analysis of Urban Landscape Therapeutic Effects with XGBoost and DiD
Malaysian Journal of Social Sciences and Humanities (MJSSH) · 14.10.2026 · 11 (10) · e004015
DOI: 10.47405/mjssh.v11i10.4015 ↗Kurz gefasst
Fachartikel in der Zeitschrift Malaysian Journal of Social Sciences and Humanities (MJSSH), online erschienen am 14.10.2026. Verfasst von Dan Wan und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Urban landscapes play a crucial role in promoting mental and physical health, yet existing methods often fail to integrate diverse data sources for assessing transformation impacts, leading to incomplete evaluations of therapeutic effects. This study addresses this gap by combining computer vision (CV) analysis of street view imagery, natural language processing (NLP) for text sentiment analysis, and environmental data (weather and air quality index) via XGBoost, followed by causal inference using difference-in-differences (DiD) to assess treatment effects of landscape changes. The aims are to develop a robust multimodal fusion model using XGBoost and apply difference-in-differences (DiD) for causal inference. Data from empirical datasets across 10 urban sites in cities like Beijing, changchun, and Nanjing (2020–2025) were analyzed, revealing a 45% increase in therapeutic indices post-transformation, with greenery visibility (35% importance) and positive sentiment (28%) as key predictors. Causal estimates confirmed significant health benefits, particularly in high-pollution areas (ATT=0.52, p
Zitieren
APA
Wan, D., Rahman, N. B. A., & Hamzah, H. B. (2026). Multimodal Causal Inference Analysis of Urban Landscape Therapeutic Effects with XGBoost and DiD. Malaysian Journal of Social Sciences and Humanities (MJSSH), 11(10), e004015. https://doi.org/10.47405/mjssh.v11i10.4015
Vancouver
Wan D, Rahman NBA, Hamzah HB. Multimodal Causal Inference Analysis of Urban Landscape Therapeutic Effects with XGBoost and DiD. Malaysian Journal of Social Sciences and Humanities (MJSSH). 2026;11(10):e004015. doi:10.47405/mjssh.v11i10.4015
BibTeX
@article{Wan2026,
author = {Wan, Dan and Rahman, Norhafizah Binti Abdul and Hamzah, Helmi Bin},
title = {Multimodal Causal Inference Analysis of Urban Landscape Therapeutic Effects with XGBoost and DiD},
journal = {Malaysian Journal of Social Sciences and Humanities (MJSSH)},
year = {2026},
volume = {11},
number = {10},
pages = {e004015},
doi = {10.47405/mjssh.v11i10.4015}
}
EUVIMED