Search

Prediction of fine particulate matter with particle size less than 2.5 µm (PM2.5) is an important component of atmospheric pollution warning and control management. In this study, we propose a deep learning model, namely, a spatiotemporal weighted neural network (STWNN), to address the challenge of poor long-term PM2.5 prediction in areas with sparse and uneven stations. The model, which is based on convolutional neural network–bidirectional long short-term memory (CNN–Bi-LSTM) and attention mechanisms and uses a geospatial data-driven approach, considers the spatiotemporal heterogeneity effec It is correct.ts of PM2.5. This approach effectively overcomes instability caused by sparse station data in forecasting daily average PM2.5 concentrations over the next week. The effectiveness of the STWNN model was evaluated using the Xinjiang Uygur Autonomous Region as the study area. Experimental results demonstrate that the STWNN exhibits higher performance (RMSE = 10.29, MAE = 6.4, R2 = 0.96, and IA = 0.81) than other models in overall prediction and seasonal clustering. Furthermore, the SHapley Additive exPlanations (SHAP) method was introduced to calculate the contribution and spatiotemporal variation of feature variables after the STWNN prediction model. The SHAP results indicate that the STWNN has significant potential in improving the performance of long-term PM2.5 prediction at the regional station level. Analyzing spatiotemporal differences in key feature variables that influence PM2.5 provides a scientific foundation for long-term pollution control and supports emergency response planning for heavy pollution events. [ABSTRACT FROM AUTHOR]

• Develop a deep learning model (GWNN) for landslide susceptibility mapping. • Compare the performance of SVM, ANN, RF, GWR and GWNN models. • The proposed model can capture the spatial heterogeneity of landslides. • Reliable landslide susceptibility map in Yichang City using the proposed model. Landslides are among the most devastating natural disasters worldwide. Landslide susceptibility mapping (LSM) is a scientific approach for assessing landslides-prone areas. Despite the use of various statistical and machine learning (ML) models to predict landslide susceptibility, few have considered the effect of spatial location on landslides. Geographically weighted regression (GWR) is a common method for modeling spatial heterogeneity, but it may not be effective in solving complex nonlinear problems. This leads to inadequate prediction accuracy for landslide susceptibility at the local scale. This study introduces a geographically weighted neural network (GWNN) model to address this problem. The GWNN model combines the advantages of GWR and neural networks (NN) to assess the effect of landslide conditioning factors on landslides. The effectiveness of the proposed method was evaluated using landslide data collected from Yichang City, Hubei Province. The results show that GWNN exhibits the higher performance (AUC = 0.788) compared with traditional ML models and GWR. Furthermore, the superior ability of GWNN is demonstrated to capture spatial heterogeneity and predict landslide susceptibility accurately at local scales. The proposed method provides a valuable reference for modeling landslide-prone areas that exhibit spatial heterogeneity. [ABSTRACT FROM AUTHOR]