A5101763807- Statistical Methods and Inference
- Face and Expression Recognition
- Bayesian Methods and Mixture Models
- Statistical Methods and Bayesian Inference
- Gene expression and cancer classification
- Machine Learning and Data Classification
- Neural Networks and Applications
- Domain Adaptation and Few-Shot Learning
- solar cell performance optimization
- Heat shock proteins research
- Ubiquitin and proteasome pathways
- Bioinformatics and Genomic Networks
- Advanced Graph Neural Networks
- Data Stream Mining Techniques
- FOXO transcription factor regulation
- Explainable Artificial Intelligence (XAI)
- Advanced Neuroimaging Techniques and Applications
- Dementia and Cognitive Impairment Research
- Recommender Systems and Techniques
- Multi-Criteria Decision Making
- Data-Driven Disease Surveillance
- Generative Adversarial Networks and Image Synthesis
- Data Visualization and Analytics
- Sensory Analysis and Statistical Methods
- Privacy-Preserving Technologies in Data
Simula Metropolitan Center for Digital Engineering
2023-2024
University of Louisiana at Lafayette
2020-2022
Nong Lam University Ho Chi Minh City
2020
Ho Chi Minh City University of Technology
2013-2018
New England College of Optometry
2007
The covariance matrix is a foundation in numerous statistical and machine-learning applications such as Principle Component Analysis, Correlation Heatmap, etc. However, missing values within datasets present formidable obstacle to accurately estimating this matrix. While imputation methods offer one avenue for addressing challenge, they often entail trade-off between computational efficiency estimation accuracy. Consequently, attention has shifted towards direct parameter estimation, given...
Abstract Forkhead‐Box Class O 4 (FOXO4) is involved in critical biological functions, but its response to EGF‐PKB/Akt signal regulation not well characterized. Here, it reported that FOXO4 levels are downregulated EGF treatment, with concurrent elevation of COP9 Signalosome subunit 6 (CSN6) and E3 ubiquitin ligase constitutive photomorphogenic 1 (COP1) levels. Mechanistic studies show CSN6 binds regulates stability through enhancing the activity COP1, COP1 directly interacts a VP motif on...
An efficient maximum power point tracking (MPPT) scheme is necessary to improve the efficiency of a solar photovoltaic (PV) panel. This paper proposes an advanced perturbation and observation (P&O) algorithm for (MPP) PV Solar cells have non-linear V-I characteristic with distinct MPP which depends on environmental factors such as temperature irradiation. In order continuously harvest from panel, it always has be operated at its MPP. The proposed P&O can reduce main drawbacks commonly...
In the global market economy, most power consulting enterprises in Vietnam have many engineering investment projects.The successful implementation of projects is a vital part their business.For to succeed, CEO and leaders those firms must make right criteria for evaluating projects.In real practice, it not an easy task because there are factors that influence success project such as time, cost, quality, satisfaction stakeholders, etc. Assessing importance these complex multi-criterion...
For many use cases, combining information from different datasets can be of interest to improve a machine learning model's performance, especially when the number samples at least one is small. However, potential challenge in such cases that features these are not identical, even though there some commonly shared among datasets. To tackle this challenge, we propose novel framework called Combine based on Imputation (ComImp). In addition, variant ComImp uses Principle Component Analysis...
For many use cases, combining information from different datasets can be of interest to improve a machine learning model's performance, especially when the number samples at least one is small. An additional challenge in such cases that features these are not identical, even though there some commonly shared among datasets. To tackle this, we propose novel framework called Combine based on Imputation (ComImp). In addition, PCA-ComImp, variant ComImp utilizes Principle Component Analysis...
Missing data is a prevalent issue that can significantly impair model performance and interpretability. This paper briefly summarizes the development of field missing with respect to Explainable Artificial Intelligence experimentally investigates effects various imputation methods on calculation Shapley values, popular technique for interpreting complex machine learning models. We compare different strategies assess their impact feature importance interaction as determined by values....
Healthcare time series data is vital for monitoring patient activity but often contains noise and missing values due to various reasons such as sensor errors or interruptions. Imputation, i.e., filling in the values, a common way deal with this issue. In study, we compare imputation methods, including Multiple Imputation Random Forest (MICE-RF) advanced deep learning approaches (SAITS, BRITS, Transformer) noisy, terms of MAE, F1-score, AUC, MCC, across rates (10 % - 80 %). Our results show...
The missing data problem has been broadly studied in the last few decades and various applications different areas such as statistics or bioinformatics. Even though many methods have developed to tackle this challenge, most of those are imputation techniques that require multiple iterations through before yielding convergence. In addition, approaches may introduce extra biases noises estimated parameters. work, we propose novel algorithms find maximum likelihood estimates (MLEs) for a...
Correlation matrix visualization is essential for understanding the relationships between variables in a dataset, but missing data can pose significant challenge estimating correlation coefficients. In this paper, we compare effects of various methods on plot, focusing two common patterns: random and monotone. We aim to provide practical strategies recommendations researchers practitioners creating analyzing plot. Our experimental results suggest that while imputation commonly used data,...
Monotone missing data is a common problem in analysis. However, imputation combined with dimensionality reduction can be computationally expensive, especially the increasing size of datasets. To address this issue, we propose Blockwise principal component analysis Imputation (BPI) framework for and monotone data. The conducts Principal Component Analysis (PCA) on observed part each block then imputes merging obtained components using chosen technique. BPI work various techniques...