Skip to content
Log in
Sponsored Have an interesting project or opportunity? → Hire me ↗
Sponsored Design Multi-Agent AI Systems Using MCP and A2A: Engineer your own Pyt Design Multi-Agent AI Systems Using MCP and A2A: Engineer your own Pyt

Author

Yi Dong

View on Semantic Scholar

2 papers indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Open access Aug 2026

Lag effects of meteorological factors on the risk of daily outpatient visits for herpangina in Guangzhou (2014–2023)

Objective Herpangina (HA) is a common acute infectious disease in children that may exhibit epidemiological patterns associated with meteorological factors. This study investigated the short-term and lagged associations of temperature and relative humidity with daily outpatient visits for HA in Guangzhou and explored patterns across sex and age groups. Methods Daily outpatient visits for HA, daily mean temperature, relative humidity, and particulate matter with an aerodynamic diameter of 2.5 micrometers or less (PM2.5) were collected from January 1, 2014 to December 31, 2023. A quasi-Poisson generalized linear model combined with distributed lag nonlinear models was used to estimate exposure-lag-response associations. Temperature, relative humidity, and PM2.5 cross-basis terms were included simultaneously, with adjustment for long-term trends, seasonality, day of week, public holidays, and the COVID-19 period. Results During the study period, 83,146 HA outpatient visits were recorded, 56.4% of which were among male patients. Daily outpatient visits showed clear seasonal variation, with larger peaks generally occurring during summer and autumn. Compared with the median temperature (24.38 °C), the cumulative RR was 0.482 (95% CI: 0.353–0.657) at the 2.5th percentile (10.77 °C) and 1.966 (95% CI: 1.714–2.254) at the 97.5th percentile (30.63 °C) over lags 0–14 days. Compared with the median relative humidity (79.15%), the cumulative RR was 0.878 (95% CI: 0.795–0.970) at the 10th percentile (62.21%) and 1.477 (95% CI: 1.297–1.682) at the 97.5th percentile (93.21%) over lags 0–20 days. High-temperature and high-humidity associations were observed across sex and age strata. The lower-humidity association was less robust in sensitivity analyses. Conclusion High temperature and high relative humidity were associated with increased daily outpatient visits for HA in Guangzhou, whereas low temperature and lower relative humidity were associated with fewer visits in the main model. The lower-humidity estimate was less stable across alternative temporal-control settings. These findings support incorporating lagged meteorological information into HA surveillance and preparedness.

Feifei Yan, Rong Xu, Yi Dong et al. · 0 citations
Preprint Aug 2026

AirFlow: Context Preserving and Multi-Rate State Modeling for Air Quality Forecasting

Accurate air quality forecasting is essential for public health and urban environmental management, but remains challenging because pollutant channels differ in periodicity and distribution drift, while their concentration trajectories contain both multi-scale dependencies and rapid changes. Recent methods have improved spatial dependency learning and meteorological covariate modeling. However, pollutant channels are still passed through the same normalization rule and temporal backbone, using a shared latent representation for channel-specific distributions and changes at different rates. To address this limitation, we propose AirFlow, a pollutant-aware dual-stream framework that operates on station multivariate observations without additional graph propagation or predefined signal decomposition. Specifically, AirFlow designs two novel blocks: (1) a statistic-guided normalization routing mechanism that selects a normalization path for each pollutant according to its 24-hour autocorrelation and distribution drift; and (2) a hierarchical dual-stream state model that combines multi-scale state space propagation with learnable response coefficients, where gated bidirectional cross-attention exchanges information and adaptively fuses the resulting representations. Experiments on real-world data from multiple cities show that AirFlow achieves the best performance in 34 of 36 metrics comparisons, with reductions of up to 11.11% root mean square error over the state-of-the-art baseline. AirFlow also requires only 0.0483M parameters and 0.0215G FLOPs, achieving high forecasting accuracy with low computational overhead.

Fan Yang, Nan Chen, Yi Dong et al. · 0 citations