Zhili Zhao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zhili Zhao has a total of 67 co-authors. Here are the top co-authors by number of shared publications:
- Ruisheng Zhang
Lanzhou University
8 shared publications Nana Zhang
School of Information Science & Engineering, Lanzhou University, Lanzhou, China
6 shared publicationsAhui Hu
School of Information Science & Engineering, Lanzhou University, Lanzhou, China
5 shared publicationsRuiyi Yan
School of Information Science and Engineering, Lanzhou University, Lanzhou, China
5 shared publicationsRuisheng Zhang
4 shared publications- Jing Jiang
Key Laboratory of Linguistic and Cultural Computing, Ministry of Education, Northwest Minzu University, Lanzhou 730030, China
3 shared publications
Latest Publications
- 22 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. PP (2162237X) , 1-15
Node Classification in GNNs: Impact of Neighborhood Label Distribution on Homophily and Heterophily
In node classification, traditional graph neural networks (GNNs) typically assume implicit homophily, indicating that intraclass nodes are likely connected. However, real-world graphs frequently exhibit heterophily, in which interclass nodes are also commonly connected. To address this challenge, recent methods have adopted approaches such as expanding local neighborhoods and employing adaptive message aggregation to enhance the GNN performance on heterophily graphs. Nevertheless, these methods are restricted by the homophily assumption and fail to effectively capture long-range dependencies (e.g., widely separated intraclass nodes) and insufficiently leverage the graph topology. This study investigates the performance differences of GNN when it is applied to both homophily and heterophily graphs and finds that the distinguishability of neighborhood label distributions (NLDs) exhibits a significant correlation with the accuracy of node classification. To assess the impact of NLD on node classification, this study proposes a novel homophily metric based on node distinguishability. Subsequently, this study introduces a new GNN model named NLD-based GNN (NLDGNN) for node classification. First, NLDGNN initializes node representations by integrating node features with node NLDs. To address long-range dependencies in heterophily graphs, NLDGNN utilizes the global label relationship matrix with low-rank characteristics for global message passing. By combining the attention scores derived from the initial node representations, NLDGNN constructs the global label relationship matrix for enhanced message passing, thereby improving the expressiveness of node representations. Experimental results indicate that NLDGNN outperforms existing GNN models on both real-world homophily and heterophily graphs. The code of this study is available at https://github.com/wanli6/NLDGNN
- 22 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. PP (2162237X) , 1-15
Node Classification in GNNs: Impact of Neighborhood Label Distribution on Homophily and Heterophily
In node classification, traditional graph neural networks (GNNs) typically assume implicit homophily, indicating that intraclass nodes are likely connected. However, real-world graphs frequently exhibit heterophily, in which interclass nodes are also commonly connected. To address this challenge, recent methods have adopted approaches such as expanding local neighborhoods and employing adaptive message aggregation to enhance the GNN performance on heterophily graphs. Nevertheless, these methods are restricted by the homophily assumption and fail to effectively capture long-range dependencies (e.g., widely separated intraclass nodes) and insufficiently leverage the graph topology. This study investigates the performance differences of GNN when it is applied to both homophily and heterophily graphs and finds that the distinguishability of neighborhood label distributions (NLDs) exhibits a significant correlation with the accuracy of node classification. To assess the impact of NLD on node classification, this study proposes a novel homophily metric based on node distinguishability. Subsequently, this study introduces a new GNN model named NLD-based GNN (NLDGNN) for node classification. First, NLDGNN initializes node representations by integrating node features with node NLDs. To address long-range dependencies in heterophily graphs, NLDGNN utilizes the global label relationship matrix with low-rank characteristics for global message passing. By combining the attention scores derived from the initial node representations, NLDGNN constructs the global label relationship matrix for enhanced message passing, thereby improving the expressiveness of node representations. Experimental results indicate that NLDGNN outperforms existing GNN models on both real-world homophily and heterophily graphs. The code of this study is available at https://github.com/wanli6/NLDGNN
- 13 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Big Data
- Vol. 12 (3) , 948-961
Modeling Complex Infectious Disease Dynamics by Distinguishing Susceptible Populations
Predicting the dynamics of infectious diseases plays an important role in their prevention and control. However, traditional epidemic models, such as susceptible–infected–recovered (SIR), susceptible–exposed–infected–recovered (SEIR), and their variants, are often inadequate for accurately capturing behavior-driven and heterogeneous transmission at scale. In this study, we propose a nonlinear dynamic model—the susceptible–low-susceptible–exposed–infected–recovered–death (SLEIRD) model—that further extends the SEIR framework. SLEIRD explicitly differentiates susceptible populations by introducing a low-susceptible group adopting protective measures, models four infection channels from susceptible/low-susceptible to exposed/infectious groups, thereby providing a behavior-aware and structurally richer compartmentalization than classical models. To validate its effectiveness, we compare SLEIRD with other classical epidemic models using real-world epidemic data, predicting confirmed cases and deaths over five, ten, fifteen, twenty, and thirty days. Experimental results show that SLEIRD outperforms SIR, improving the root mean squared error (RMSE) for daily confirmed cases and deaths by 6.0% –46.3% and 31.1% –62.3%, respectively; compared with SEIR, SLEIRD reduces the average RMSE for daily confirmed cases and deaths by 6.8% –56.6% and 59.8% –79.6%, respectively. These findings indicate that SLEIRD more accurately simulates complex spread dynamics and offers a scalable, behavior-aware tool for data-driven epidemic forecasting with practical value for disease control and prevention.
- 13 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Big Data
- Vol. 12 (3) , 948-961
Modeling Complex Infectious Disease Dynamics by Distinguishing Susceptible Populations
Predicting the dynamics of infectious diseases plays an important role in their prevention and control. However, traditional epidemic models, such as susceptible–infected–recovered (SIR), susceptible–exposed–infected–recovered (SEIR), and their variants, are often inadequate for accurately capturing behavior-driven and heterogeneous transmission at scale. In this study, we propose a nonlinear dynamic model—the susceptible–low-susceptible–exposed–infected–recovered–death (SLEIRD) model—that further extends the SEIR framework. SLEIRD explicitly differentiates susceptible populations by introducing a low-susceptible group adopting protective measures, models four infection channels from susceptible/low-susceptible to exposed/infectious groups, thereby providing a behavior-aware and structurally richer compartmentalization than classical models. To validate its effectiveness, we compare SLEIRD with other classical epidemic models using real-world epidemic data, predicting confirmed cases and deaths over five, ten, fifteen, twenty, and thirty days. Experimental results show that SLEIRD outperforms SIR, improving the root mean squared error (RMSE) for daily confirmed cases and deaths by 6.0% –46.3% and 31.1% –62.3%, respectively; compared with SEIR, SLEIRD reduces the average RMSE for daily confirmed cases and deaths by 6.8% –56.6% and 59.8% –79.6%, respectively. These findings indicate that SLEIRD more accurately simulates complex spread dynamics and offers a scalable, behavior-aware tool for data-driven epidemic forecasting with practical value for disease control and prevention.
- 4 May 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 287
Node influence-based label propagation for community detection using both topology and attributes
- 4 May 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 287
Node influence-based label propagation for community detection using both topology and attributes
- 10 April 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 281
Maximizing influence by combining influential node identification and overlapping influence reduction
- 10 April 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 281
Maximizing influence by combining influential node identification and overlapping influence reduction
- Open Access
A Modularity-Enhanced Echo State Network for Nonlinear Wind Energy Predicting
With the rapid growth of wind power generation, accurate wind energy prediction has emerged as a critical challenge, particularly due to the highly nonlinear nature of wind speed data. This paper proposes a modularized Echo State Network (MESN) model to improve wind energy forecasting. To enhance generalization, the wind speed data is first decomposed into time series components, and Modes-cluster is employed to extract trend patterns and pre-train the ESN output layer. Furthermore, Turbines-cluster groups wind turbines based on their wind speed and energy characteristics, enabling turbines within the same category to share the ESN output matrix for prediction. An output integration module is then introduced to aggregate the predicted results, while the modular design ensures efficient task allocation across different modules. Comparative experiments with other neural network models demonstrate the effectiveness of the proposed approach, showing that the statistical RMSE of parameter error is reduced by an average factor of 2.08 compared to traditional neural network models.
Open AccessA Modularity-Enhanced Echo State Network for Nonlinear Wind Energy Predicting
With the rapid growth of wind power generation, accurate wind energy prediction has emerged as a critical challenge, particularly due to the highly nonlinear nature of wind speed data. This paper proposes a modularized Echo State Network (MESN) model to improve wind energy forecasting. To enhance generalization, the wind speed data is first decomposed into time series components, and Modes-cluster is employed to extract trend patterns and pre-train the ESN output layer. Furthermore, Turbines-cluster groups wind turbines based on their wind speed and energy characteristics, enabling turbines within the same category to share the ESN output matrix for prediction. An output integration module is then introduced to aggregate the predicted results, while the modular design ensures efficient task allocation across different modules. Comparative experiments with other neural network models demonstrate the effectiveness of the proposed approach, showing that the statistical RMSE of parameter error is reduced by an average factor of 2.08 compared to traditional neural network models.
- Open Access
- 18 March 2025
- Published by MDPI AG in Preprints.org
A Modularity-Enhanced Echo State Network for Nonlinear Wind Energy Predicting
With the rapid growth of wind power generation, accurate wind energy prediction has emerged as a critical challenge, particularly due to the highly nonlinear nature of wind speed data. This paper proposes a modularized Echo State Network (MESN) model to improve wind energy forecasting. To enhance generalization, the wind speed data is first decomposed into time series components, and Modes-cluster is employed to extract trend patterns and pre-train the ESN output layer. Furthermore, Turbines-cluster groups wind turbines based on their wind speed and energy characteristics, enabling turbines within the same category to share the ESN output matrix for prediction. An output integration module is then introduced to aggregate the predicted results, while the modular design ensures efficient task allocation across different modules. Comparative experiments with other neural network models demonstrate the effectiveness of the proposed approach, showing that the statistical RMSE of parameter error is reduced by an average factor of 2.08 compared to traditional neural network models.
- 18 March 2025
- Published by MDPI AG in Preprints.org
Open AccessA Modularity-Enhanced Echo State Network for Nonlinear Wind Energy Predicting
With the rapid growth of wind power generation, accurate wind energy prediction has emerged as a critical challenge, particularly due to the highly nonlinear nature of wind speed data. This paper proposes a modularized Echo State Network (MESN) model to improve wind energy forecasting. To enhance generalization, the wind speed data is first decomposed into time series components, and Modes-cluster is employed to extract trend patterns and pre-train the ESN output layer. Furthermore, Turbines-cluster groups wind turbines based on their wind speed and energy characteristics, enabling turbines within the same category to share the ESN output matrix for prediction. An output integration module is then introduced to aggregate the predicted results, while the modular design ensures efficient task allocation across different modules. Comparative experiments with other neural network models demonstrate the effectiveness of the proposed approach, showing that the statistical RMSE of parameter error is reduced by an average factor of 2.08 compared to traditional neural network models.
- 6 February 2025
- Published by Elsevier in Knowledge-Based Systems
- Vol. 311
Personalized recommendation by integrating a neural topic model and Bayesian personalized ranking
- 6 February 2025
- Published by Elsevier in Knowledge-Based Systems
- Vol. 311
Personalized recommendation by integrating a neural topic model and Bayesian personalized ranking
- 7 July 2024
- Published by Elsevier in Expert Systems with Applications
- Vol. 255
Integrating topology and content equally in non-negative matrix factorization for community detection
- 7 July 2024
- Published by Elsevier in Expert Systems with Applications
- Vol. 255
Integrating topology and content equally in non-negative matrix factorization for community detection
- 5 June 2024
- Published by Elsevier in Knowledge-Based Systems
- Vol. 299
Mining node attributes for link prediction with a non-negative matrix factorization-based approach
- 5 June 2024
- Published by Elsevier in Knowledge-Based Systems
- Vol. 299
Mining node attributes for link prediction with a non-negative matrix factorization-based approach
- 27 May 2024
- Published by Elsevier in Physica A: Statistical Mechanics and its Applications
- Vol. 646
Mining odd-length paths for link prediction in bipartite networks
- 27 May 2024
- Published by Elsevier in Physica A: Statistical Mechanics and its Applications
- Vol. 646
Mining odd-length paths for link prediction in bipartite networks