Ajim Uddin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Ajim Uddin has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
Xinyuan Tao
Martin Tuchman School of Management, New Jersey Institute of Technology , Newark, NJ 07102
4 shared publicationsDantong Yu
Martin Tuchman School of Management, New Jersey Institute of Technology , Newark, NJ 07102
3 shared publications- Dantong Yu
New Jersey Institute of Technology, US
3 shared publications - Dan Zhou
New Jersey Institute of Technology, United States
2 shared publications Mahmudul Haque
2 shared publicationsBhaskar Goswami
New Jersey Institute of Technology, Newark, USA
1 shared publication
Latest Publications
- Open Access
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 2915-2924
Neural Instrumented Factorization: Learning Dynamic Asset Pricing Factors and Loadings through Characteristics Control
- 10 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 2915-2924
Open AccessNeural Instrumented Factorization: Learning Dynamic Asset Pricing Factors and Loadings through Characteristics Control
- Open Access
- 24 July 2025
- Published by Springer Nature in Annals of Operations Research
- Vol. 357 (1) , 223-257
Significance of predictors: revisiting stock return predictions using explainable AI
In this paper, we re-examine 166 previously identified asset pricing characteristics and their ability to successfully predict stock returns. We use Explainable Artificial Intelligence to rank these return predictors based on their importance in various asset pricing model settings. Our findings suggest that ensemble and deep learning-based models have an advantage in providing generalized predictions across different return measures. Using SHapley Additive exPlanations, we also find that momentum and trading-based features possess higher predictive power in estimating asset returns. The long-short portfolio analysis reveals that key return predictors exhibit substantial economic significance, reflected in the large differences in out-of-sample $$R^2$$ . These findings remain robust across various models and persist even after controlling for characteristics-based predictors.- 24 July 2025
- Published by Springer Nature in Annals of Operations Research
- Vol. 357 (1) , 223-257
Open AccessSignificance of predictors: revisiting stock return predictions using explainable AI
In this paper, we re-examine 166 previously identified asset pricing characteristics and their ability to successfully predict stock returns. We use Explainable Artificial Intelligence to rank these return predictors based on their importance in various asset pricing model settings. Our findings suggest that ensemble and deep learning-based models have an advantage in providing generalized predictions across different return measures. Using SHapley Additive exPlanations, we also find that momentum and trading-based features possess higher predictive power in estimating asset returns. The long-short portfolio analysis reveals that key return predictors exhibit substantial economic significance, reflected in the large differences in out-of-sample $$R^2$$ . These findings remain robust across various models and persist even after controlling for characteristics-based predictors. - Open Access
- 14 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 388-396
DySTAGE: Dynamic Graph Representation Learning for Asset Pricing via Spatio-Temporal Attention and Graph Encodings
Current GNN-based asset price prediction models often focus on a fixed group of assets and their static relationships within the financial network. However, this approach overlooks the reality that the composition of asset pools and their interrelationships evolves over time, necessitating the development of a flexible framework capable of adapting to this dynamism. Accordingly, we propose DySTAGE, a framework with a universal formulation that transforms asset pricing time series into dynamic graphs, accommodating asset addition, deletion, and changes in correlations. Our framework includes a graph learning model specifically designed for this purpose. In our framework, assets at various historical time steps are structured as a sequence of dynamic graphs, where connections between assets reflect their long-term correlations. DySTAGE effectively captures both topological and temporal patterns. The Topological Module deploys Asset Influence Attention to learn global interrelationships among assets, further enhanced by Asset-wise Importance Encoding, Pair-wise Spatial Encoding, and Edge-wise Correlation Encoding. Meanwhile, the Temporal Module encapsulates node representations across the temporal dimension via the attention mechanism. We validate our approach through extensive experiments using three different real-world stock pricing data, demonstrating that DySTAGE surpasses popular benchmarks in return prediction, and offers profitable investment strategies. The code is publicly available under NJIT FinTech Lab’s GitHub1.
- 14 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 388-396
Open AccessDySTAGE: Dynamic Graph Representation Learning for Asset Pricing via Spatio-Temporal Attention and Graph Encodings
Current GNN-based asset price prediction models often focus on a fixed group of assets and their static relationships within the financial network. However, this approach overlooks the reality that the composition of asset pools and their interrelationships evolves over time, necessitating the development of a flexible framework capable of adapting to this dynamism. Accordingly, we propose DySTAGE, a framework with a universal formulation that transforms asset pricing time series into dynamic graphs, accommodating asset addition, deletion, and changes in correlations. Our framework includes a graph learning model specifically designed for this purpose. In our framework, assets at various historical time steps are structured as a sequence of dynamic graphs, where connections between assets reflect their long-term correlations. DySTAGE effectively captures both topological and temporal patterns. The Topological Module deploys Asset Influence Attention to learn global interrelationships among assets, further enhanced by Asset-wise Importance Encoding, Pair-wise Spatial Encoding, and Edge-wise Correlation Encoding. Meanwhile, the Temporal Module encapsulates node representations across the temporal dimension via the attention mechanism. We validate our approach through extensive experiments using three different real-world stock pricing data, demonstrating that DySTAGE surpasses popular benchmarks in return prediction, and offers profitable investment strategies. The code is publicly available under NJIT FinTech Lab’s GitHub1.
- 14 May 2024
- Published by Oxford University Press (OUP) in Journal of Financial Econometrics
- Vol. 22 (5) , 1616-1655
The Network Factor of Equity Pricing: A Signed Graph Laplacian Approach
The connections among firms exhibit heterogeneity, complexity, and dynamism, posing a challenge for traditional unsigned network models. This article proposes a signed graph Laplacian approach to construct a dynamic network index (DNI), quantifying the aggregate changes in the market network over time. A larger DNI indicates more significant changes in firms’ interconnectedness and in the market network structure. Firms with higher sensitivity to DNI exhibit lower expected returns. Incorporating DNI into conventional asset pricing models improves return predictability. Results are robust for multiple estimators, various factor models, and different selections of test assets. Our findings suggest that the network factor generates a significant equity risk premium.
- 14 May 2024
- Published by Oxford University Press (OUP) in Journal of Financial Econometrics
- Vol. 22 (5) , 1616-1655
The Network Factor of Equity Pricing: A Signed Graph Laplacian Approach
The connections among firms exhibit heterogeneity, complexity, and dynamism, posing a challenge for traditional unsigned network models. This article proposes a signed graph Laplacian approach to construct a dynamic network index (DNI), quantifying the aggregate changes in the market network over time. A larger DNI indicates more significant changes in firms’ interconnectedness and in the market network structure. Firms with higher sensitivity to DNI exhibit lower expected returns. Incorporating DNI into conventional asset pricing models improves return predictability. Results are robust for multiple estimators, various factor models, and different selections of test assets. Our findings suggest that the network factor generates a significant equity risk premium.
- Open Access
- 25 November 2023
- Published by Association for Computing Machinery (ACM)
- Vol. 16, 516-523
NMTucker: Non-linear Matryoshka Tucker Decomposition for Financial Time Series Imputation
- 25 November 2023
- Published by Association for Computing Machinery (ACM)
- Vol. 16, 516-523
Open AccessNMTucker: Non-linear Matryoshka Tucker Decomposition for Financial Time Series Imputation
- Open Access
- 25 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 409-417
A Fast Non-Linear Coupled Tensor Completion Algorithm for Financial Data Integration and Imputation
Missing data imputation is crucial in finance to ensure accurate financial analysis, risk management, investment strategies, and other financial applications. Recently, tensor factorization and completion have gained momentum in many finance data imputation applications, primarily due to recent breakthroughs in applying deep neural networks for nonlinear tensor analysis. However, one limitation of these approaches is that they are prone to overfitting sparse tensors that contain only a small number of observations. This paper focuses on learning highly reliable embedding for the tensor imputation problem and applies orthogonal regularizations for tensor factorization, reconstruction, and completion. The proposed neural network architecture for sparse tensors, called “RegTensor”, includes multiple components: an embedding learning module for each tensor order, MLP (multilayer perception) to model nonlinear interactions among embeddings, and a regularization module to minimize overfitting problems due to the large tensor rank. Our algorithm is efficient in factorizing both single and multiple tensors (coupled tensor factorization) without incurring high training and optimization costs. We have applied this algorithm in a variety of practical scenarios, including the imputation of bond characteristics and financial analyst EPS forecast data. Experimental results demonstrate its superiority with significant performance improvements: 40%-74% better than linear tensor completion models and 2%-52% better than the state-of-the-art nonlinear models.
- 25 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 409-417
Open AccessA Fast Non-Linear Coupled Tensor Completion Algorithm for Financial Data Integration and Imputation
Missing data imputation is crucial in finance to ensure accurate financial analysis, risk management, investment strategies, and other financial applications. Recently, tensor factorization and completion have gained momentum in many finance data imputation applications, primarily due to recent breakthroughs in applying deep neural networks for nonlinear tensor analysis. However, one limitation of these approaches is that they are prone to overfitting sparse tensors that contain only a small number of observations. This paper focuses on learning highly reliable embedding for the tensor imputation problem and applies orthogonal regularizations for tensor factorization, reconstruction, and completion. The proposed neural network architecture for sparse tensors, called “RegTensor”, includes multiple components: an embedding learning module for each tensor order, MLP (multilayer perception) to model nonlinear interactions among embeddings, and a regularization module to minimize overfitting problems due to the large tensor rank. Our algorithm is efficient in factorizing both single and multiple tensors (coupled tensor factorization) without incurring high training and optimization costs. We have applied this algorithm in a variety of practical scenarios, including the imputation of bond characteristics and financial analyst EPS forecast data. Experimental results demonstrate its superiority with significant performance improvements: 40%-74% better than linear tensor completion models and 2%-52% better than the state-of-the-art nonlinear models.
- Open Access
- 25 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 235-243
The Network of Mutual Funds: A Dynamic Heterogeneous Graph Neural Network for Estimating Mutual Funds Performance
Mutual funds are interconnected to each other through multiple types of links, including but not limited to co-investment, advisors, firms, and managers. These connections enable information flow among network entities, influence investment decisions, and ultimately impact mutual fund managers’ performance. In this paper, we propose a dynamic graph neural network approach to model these heterogeneous relationships and their contributions to mutual fund performance. Using the graph attention mechanism, our model learns latent embedding for mutual funds and their invested assets dynamically in each month and then uses the embedding to estimate future returns. Empirical analysis confirms that the proposed method outperforms the state-of-the-art DeepWalk model by 10%. Furthermore, this study also reveals the importance of networks in mutual fund performance. The inclusion of network connections in a feedforward machine learning model significantly increases the performance of the model by 118%. Finally, portfolio analysis and regression estimation on next month’s excess return show that the proposed approach has a significant economic contribution over current benchmark approaches.
- 25 November 2023
- Published by Association for Computing Machinery (ACM)
- p. 235-243
Open AccessThe Network of Mutual Funds: A Dynamic Heterogeneous Graph Neural Network for Estimating Mutual Funds Performance
Mutual funds are interconnected to each other through multiple types of links, including but not limited to co-investment, advisors, firms, and managers. These connections enable information flow among network entities, influence investment decisions, and ultimately impact mutual fund managers’ performance. In this paper, we propose a dynamic graph neural network approach to model these heterogeneous relationships and their contributions to mutual fund performance. Using the graph attention mechanism, our model learns latent embedding for mutual funds and their invested assets dynamically in each month and then uses the embedding to estimate future returns. Empirical analysis confirms that the proposed method outperforms the state-of-the-art DeepWalk model by 10%. Furthermore, this study also reveals the importance of networks in mutual fund performance. The inclusion of network connections in a feedforward machine learning model significantly increases the performance of the model by 118%. Finally, portfolio analysis and regression estimation on next month’s excess return show that the proposed approach has a significant economic contribution over current benchmark approaches.
- 2 October 2023
- Published by Elsevier in Global Finance Journal
- Vol. 58
Attention based dynamic graph neural network for asset pricing
- 2 October 2023
- Published by Elsevier in Global Finance Journal
- Vol. 58
Attention based dynamic graph neural network for asset pricing
- Open Access
- 1 January 2023
- Published by Elsevier in SSRN Electronic Journal
Network Centrality, Leadership, and Institutional Investors Portfolio Performance
This paper proposes a novel (fund-equity) bipartite network projection model to capture the network centrality of fund managers through their equity connections. The proposed alpha centrality demonstrates superiority in tracking the network evolution over time. This data-driven centrality allows us to empirically examine the importance of managers' centrality on their investment performance. We document a positive relation between managers' network centrality and investment performance where central managers on average earn 6.4% higher returns per year. We also uncover an inverted U-shaped pattern in the herding behaviors of managers. Results are consistent for different time periods and robust after multiple controls.
- 1 January 2023
- Published by Elsevier in SSRN Electronic Journal
Open AccessNetwork Centrality, Leadership, and Institutional Investors Portfolio Performance
This paper proposes a novel (fund-equity) bipartite network projection model to capture the network centrality of fund managers through their equity connections. The proposed alpha centrality demonstrates superiority in tracking the network evolution over time. This data-driven centrality allows us to empirically examine the importance of managers' centrality on their investment performance. We document a positive relation between managers' network centrality and investment performance where central managers on average earn 6.4% higher returns per year. We also uncover an inverted U-shaped pattern in the herding behaviors of managers. Results are consistent for different time periods and robust after multiple controls.
- 1 May 2022
- Published by Elsevier in Finance Research Letters
- Vol. 46
The network structure of overnight index swap rates
- 1 May 2022
- Published by Elsevier in Finance Research Letters
- Vol. 46
The network structure of overnight index swap rates