Hung Tran-Nam

Hung Tran-Nam

Auto-generated profile, based on publicly available information

Hung Tran-Nam

Affiliations Laboratory for Applied and Industrial Mathematics, Institute for Computational Science and Artificial Intelligence, Van Lang University, Ho Chi Minh City, Vietnam, Faculty of Fundamental Sciences, Van Lang University, Ho Chi Minh City, Vietnam

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
8 June 2026 in PLOS ONE
Verified
Journal Article
Open access
PDF file
HTHung Tran-Nam
Hung Tran-Nam
HCHa Che-Ngoc
Ha Che-Ngoc
Clustering uncertain data is a fundamental problem in data mining. Imbalance among uncertain objects significantly degrades clustering performance, as minority clusters are repeatedly overshadowed by dominant ones. Consequently, existing clustering techniques often fail due to initialisation biases and inadequate similarity modelling. This paper proposes a novel algorithm, the B lack-winged K ite I mproved F uzzy clustering for probability density F unctions (BKIFF), which combines an optimisation-based initialisation strategy with an enhanced fuzzy clustering framework. Specifically, BKIFF incorporates the Hellinger distance into the clustering objective to more reliably capture similarities between probability density functions (pdfs), and introduces improved membership updating and prototype estimation mechanisms tailored for uncertain and imbalanced data formulated as I mproved F uzzy clustering for probability density F unctions (IFF) while theoretical convergence is established. In addition, the algorithm employs Black-winged Kite Optimisation (BKO) to enhance prototype selection, improving clustering stability and convergence. As a result, comprehensive experiments with synthetic Gaussian probability distributions, skewed pdfs, and real-world image datasets demonstrate that BKIFF consistently outperforms baseline methods such as FCF, FCF- ℒ 1 , K MEANS, and Self-Updating. Across all three examples, BKIFF achieves near-perfect ARI , improving from near-zero values in highly imbalanced cases {20,50,80,100} by approximately 30–35% in moderate settings, while increasing NMI by about 25–95%. Additionally, it reduces computational time by approximately 95–99% compared to baseline methods. In conclusion, BKIFF demonstrates superior performance and opens up new possibilities for applications in medical diagnostics, ecological analysis, and high-dimensional uncertain data mining, particularly in imbalanced environments.

19 May 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
HTHung Tran-Nam
Hung Tran-Nam
TNThao Nguyen-Trang
Thao Nguyen-Trang
PNPhong Thanh Nguyen
Phong Thanh Nguyen
NPNhat Truong Pham
Nhat Truong Pham
AQAnh Nguyen Quoc
Anh Nguyen Quoc
DADuong Tran Anh
Duong Tran Anh
Drought conditions in the Mekong River Basin (MRB) have been exacerbated by increasing hydroclimatic variability, posing significant challenges to water resources, agricultural productivity, and ecosystem stability. Although the application of deep learning techniques in drought forecasting is on the rise, a thorough understanding of the interactions among meteorological, hydrological, and agricultural drought processes across various temporal scales remains limited, particularly in the MRB. This study introduces a comprehensive multi-index drought forecasting framework that integrates causal inference, statistical trend analysis, explainable artificial intelligence, and a comparative analysis of eight deep learning architectures. It examines three standardised drought indices, including the Standardised Precipitation Index (SPI), Standardised Runoff Index (SRI), and Standardised Soil Water Index (SSWI), across short-, mid-, and long-term accumulation scales at six mainstream hydrological stations from 1982 to 2019. Forecasting experiments were conducted to assess one- and two-year prediction horizons under consistent training conditions. The results of model benchmarking indicated that forecasting performance was more influenced by drought type and station characteristics than by the prediction horizon or lookback duration. Notably, the PatchTST, RNN and KAN models consistently demonstrated the strongest performance across multiple drought indices and timescales. The explainability analysis highlighted that the models exhibiting high performance attributed outcomes primarily to early antecedent conditions and specific temporal intervals linked to drought persistence. These findings underscore that effective drought forecasting in the MRB fundamentally relies on the interplay between drought memory, hydrological variability, and the ability of forecasting architectures to capture long-range temporal dependencies.

15 August 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
KTKhoa Tran
Khoa Tran
TLTri Le
Tri Le
HTHung-Cuong Trinh
Hung-Cuong Trinh
HTHung Tran-Nam
Hung Tran-Nam
Accurate remaining discharge time (RDT) prediction is challenging in real-world battery applications because future load profiles are unknown and highly dynamic. To address the uncertainty of continuous RDT regression, this paper introduces Relative Discharge Stage (RDS), a battery-management indicator that represents the remaining discharge condition using five interpretable classes: Normal, Good, Moderate, Low, and Recharge Required. Unlike state of charge (SOC), which reflects the current charge level, RDS characterizes the remaining discharge process without requiring future-current information during inference. A physics-informed RDS classification framework is proposed, combining SOC estimation with lightweight temporal learning. The SOC-estimation component includes second-order ECM state and terminal-voltage prediction, hysteresis and OCV temperature correction, core-temperature estimation, and AEKF state correction, supported by OCV evaluation, online STC-ECM parameter adaptation, and pretrained neural residual-voltage correction. The measured current, terminal voltage, surface temperature, and estimated SOC are arranged into a sliding observation window and processed by a lightweight temporal convolutional network. Experiments on two public lithium-ion battery datasets demonstrate robust RDS classification, with accuracy exceeding 80% under varying load and thermal conditions.

25 October 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
HTHung Tran-Nam
Hung Tran-Nam
TNThao Nguyen-Trang
Thao Nguyen-Trang
PNPhong Thanh Nguyen
Phong Thanh Nguyen
NPNhat Truong Pham
Nhat Truong Pham
AQAnh Nguyen Quoc
Anh Nguyen Quoc
Duong Tran AnhDuong Tran Anh
Duong Tran Anh
While data-driven forecasting has advanced rapidly, there remains a need for systematic benchmarking across heterogeneous model architectures, spatially distributed stations, and interacting drought indices. In this study, we develop a unified multi-index forecasting framework for meteorological (SPI), hydrological (SRI), and agricultural (SSWI) drought conditions across six stations in the MRB. The framework evaluates 15 artificial intelligence architectures spanning statistical learning (SL), machine learning (ML), and deep learning (DL), together with four progressive feature-engineering configurations and two multi-output forecasting strategies. The evaluation is conducted using strictly chronological data partitioning to preserve the temporal ordering of observations and to avoid the use of future information during model development. Results indicate that tree-based ensemble models, particularly ExtraTrees and LightGBM, generally achieve lower prediction errors than the evaluated deep learning architectures under the available monthly sample size, with a large standardized Cohen's effect size of 0.87. Temporal feature engineering provides the most consistent improvement, whereas the tested spatially distributed station information contributes comparatively smaller gains. Granger-based analyses identify statistically supported temporal dependencies among drought indicators, while SHAP analysis reveals physically plausible contributions from antecedent hydrometeorological conditions, particularly water level and soil-moisture deficits. Overall, the results provide a systematic methodological benchmark for multi-index drought forecasting and a practical reference for future drought early-warning applications in transboundary river basins.