DR-NTU (Digital Repository of NTU)
Not a member yet
116018 research outputs found
Sort by
Childhood harsh discipline and behavioral problems: the mediating roles of attachment dimensions
No full textBackground:
Parental harsh discipline (i.e., psychological aggression and physical discipline) has been consistently associated with greater child behavioral problems. Yet the mechanism underlying this association remains unclear. Harsh discipline may not have similar negative implications in cultures where it is prevalent.
Objective:
This study examined the direct associations between childhood harsh discipline and behavioral problems, as well as the mediating roles of attachment dimensions (i.e., trust, communication, alienation) in these relationships.
Participants and setting:
449 Singaporean young adults completed the survey (Mage = 22.57, SDage = 1.77, 52% female).
Methods:
Participants reported on their childhood discipline, current attachment towards their parent, and current behavioral (i.e., externalizing and internalizing) problems. Path analyses were conducted separately for mothers and fathers.
Results:
Maternal psychological aggression and parental severe physical discipline were directly related to greater externalizing problems. Parental psychological aggression and maternal severe physical discipline were indirectly associated with greater behavioral problems through alienation. The relationship between paternal psychological aggression and externalizing problems was also mediated by poor communication. Null finding for minor physical discipline was observed.
Conclusions:
Psychological aggression and severe physical discipline are related to greater behavioral problems, even in a cultural context where strict discipline is deemed acceptable. Minor physical discipline may have less adverse consequences in such a context. Parental alienation and paternal poor communication were the intervening mechanisms through which psychological aggression and severe physical discipline are related to greater behavioral problems. These findings highlight the need to encourage non-violent disciplinary strategies to facilitate optimal child development.Ministry of Education (MOE)Published versionThis research was supported by grants from the Singapore Ministry of Education Academic Research Fund Tier 1 (RG126/23; RG39/22; RG42/20) and Yong Loo Lin School of Medicine, National University of Singapore (NUHSRO/2021/093/NUSMed/13/LOA) awarded to Peipei Setoh
Application of extreme value theory in the financial industry
No full textThis project explores the application of Extreme Value Theory (EVT) for developing
risk management frameworks in the cryptocurrency market a domain characterized
by high volatility and frequent extreme price movements. Recognizing the
limitations of traditional risk models in capturing rare but impactful market events -
this study aims to employ the Generalized Pareto Distribution (GPD) to model tail
risks effectively. By applying EVT to financial datasets the research identifies
extreme events and quantifies their likelihood and potential impact. The proposed
framework enhances the accuracy of Value at Risk (VaR) and Expected Shortfall
(ES) calculations providing financial institutions with more reliable risk assessments.
The findings contribute to improved decision-making in portfolio management and
stress testing. This research also offers actionable insights for financial practitioners
seeking to mitigate extreme market risks through more resilient risk management
strategies and practices.Bachelor's degre
Explainable multi-step time series forecasting model
No full textThis report presents an explainable multi-step time series forecasting model called the MSG-TCN. This model integrates multi-scale dilated convolutions (M), self-attention mechanisms (S) and gated activation units (G) on a Temporal Convolutional Network (TCN) network to effectively capture both local and global temporal patterns. We benchmark the MSG-TCN against four baselines, namely Autoregressive Integrated Moving Average (ARIMA), Long Short Term Memory (LSTM), a standard TCN, and the Galformer transformer model. For datasets, we used three major stock market indices (Standard & Poor’s 500 (S&P 500), Dow Jones Industrial Average (DJIA) and Nasdaq Composite Index (IXIC)). Experimental evaluations cover various forecast horizons (1-day, 10-day and 20-day), measuring performance via Mean Absolute Percentage Error (MAPE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE) and the coefficient of determination (2). Results demonstrate that MSG-TCN consistently attains comparable accuracy to state-of-the-art baselines while retaining a transparent structure amenable to interpretability. We also integrated local interpretable Shapley additive explanation methods into MSG-TCN, allowing us to visualize the key features that influence the model’s decisions and improve its interpretability. By merging precise multi-step forecasting with enhanced interpretability, our project delivers a robust, scalable solution for high-stakes decision-making while driving continued improvements in the financial forecasting domain.Bachelor's degre
Multimodal medical data analysis using deep neural network
No full textThe integration of multiple modalities has become a promising approach in the medical field to address limitations of single-sourced data. This project explores multimodal medical data analysis using deep neural networks, to improve the classification of Alzheimer’s Disease (AD), Mild Cognitive Impairment (MCI), and healthy subjects (CN). Acquiring data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI), this study incorporates 2 distinct modalities: T1-weighted MRI and T2 FLAIR-weighted MRI. The project aims to enhance classification accuracy beyond what individual modalities can achieve.
The proposed model in this project is a pretrained 3D Convolutional Neural Network (3D-CNN), 3D MobileNetV2 1.0x using pretrained weights from Kinetics-600 video dataset. Various fusion techniques will be explored to provide a comprehensive understanding on the capabilities of multimodal data. This includes Early, Late, Intermediate and the project’s own Novelty Fusion. The best performing classification accuracy was 0.898, achieved by Novelty Fusion (Early + Intermediate). Comparatively, this outperformed single-modality scores, where T1-weighted MRI obtained 0.4926 and T2-weighted MRI obtained 0.5519. The significant improvement in accuracy showcased the effectiveness of multimodal integration, in particular for more complex fusion variations.
While this model only used 2 MRI modalities, such a framework is highly adaptable to other data types such as positron emission tomography (PET), electrocardiogram (ECG) and non-imaging modalities like clinical assessments and audio-based data. This flexibility lays the groundwork for building medical diagnostic systems that utilise a wide array of patient information. It is hoped that the insights gained from this study will contribute to future medical studies, aiding in more timely, accurate and precise patient care.Bachelor's degre
Online secure payment for membership system
No full textTechnology is rapidly advancing at a fast speed, powering industries and economies around the world, yet access to timely and up-to-date information is not equitable. In China, strict internet censorship policies, commonly referred to as the "Great Firewall," restrict access to international platforms, academic research, open-source communities, and technical forums.
Censorship does not just create barriers to the sharing of knowledge but also to innovation, collaboration, and competitiveness in fields like artificial intelligence, cybersecurity, and software development.
As a remedy for this issue, this project proposes an open-access resource-learning website with complete and free access to research papers, tutorials, and forums on emerging technologies. By bridging the gap between Chinese technology enthusiasts and the rest of the world's technology landscape, the site will foster learning, innovation, and collaboration. As the rapidly evolving digital age, open exchange of knowledge is the most important thing to stay competitive, and this project intends to empower individuals and businesses with the ability to contribute to and access the world's technological space without any limitations.Bachelor's degre
Multi-modal large language models for ophthalmology triage
No full textThe increasing prevalence of ocular diseases worldwide and the resultant burden on
healthcare systems has underscored the need for accurate, consistent, and scalable
triage systems. This study explores the application of large vision-language models
(VLMs) in improving diagnostic accuracy and, by extension, ophthalmic triage accu-
racy. We propose a comprehensive framework that integrates structured clinical text
generation from unstructured notes, supported by hallucination detection to ensure
input reliability. To robustly evaluate diagnostic performance, we introduce a graph-
based method that leverages a Directed Acyclic Graph (DAG) of medical concepts
to compute dissimilarity scores between predicted and ground-truth diagnoses. This
enables a more nuanced assessment of model output beyond exact label matching. We
conduct a multimodal evaluation using various ophthalmic imaging modalities, com-
paring text-only and image-assisted diagnoses. Our findings show that while image
inputs can significantly enhance diagnostic accuracy for certain conditions, they may
degrade performance in others—highlighting the need for context-aware integration of
visual data. This work establishes a foundation for more interpretable and clinically
aligned triage support systems powered by multimodal large language models (LLMs).Bachelor's degre
Generation of beam profiles from chip-to-free-space coupling using deep neural network
No full textThree-dimensional (3D) simulation of light propagating from silicon photonics (SiPh) gratings to free space using the finite-difference time-domain (FDTD) technique can be time-consuming, with simulation times reaching up to 12 hours. To address this issue, we propose a deep neural network (DNN) modeling technique as a rapid, alternative method for visualizing light propagation. In this work, a DNN model is developed to generate beam profiles in chip-to-free-space coupling of SiPh gratings. The DNN model was trained only with 10% of the beam profiles from the FDTD-simulated 3D electric field (E-field), and used to generate the remaining 90% of the beam profiles. The total training time for the DNN is 17 minutes, while the time taken for E-field prediction and beam profile generation is less than 10 minutes. Comparing the FDTD-simulated and DNN-generated beam profiles, the average percentage error (APE) for beam waists is 4.69%, and the APE for the maximum E-field is 8.80%. When varying the architecture of the DNN model, prediction accuracies of > 0.94 and > 0.87 are obtained for beam waists and maximum E-fields, respectively.Ministry of Education (MOE)Submitted/Accepted versionThis work was supported by the Ministry of Education of Singapore AcRF Tier 2 (T2EP50121-0002 (MOE-000180-01)) and AcRF Tier 1 (RG135/23, RT3/23)
VR games in metaverse IV
No full textIn recent years, Virtual Reality(VR) technology has steadily gained popularity. While there was not much of a market for VR applications in the early days, as the technology became more mainstream, more games started to be developed for VR. VR remains a niche market compared to other platforms like PC and consoles like Switch or PS5. However, given its current rapid growth, it would be beneficial to explore the creation of different kinds of VR games in order to meet the rising demand. This project focuses on the development of a cooperative 3D block puzzle game for 2 players, where they have to communicate with each other to solve multiple puzzles of varying difficulty levels.Bachelor's degre
SPL fandoms: a study into persevering fandoms
No full textThe Singapore Premier League’s (SPL) popularity has been waning throughout the years. Once a popular past-time for football fans throughout Singapore, fan support for the league has since dwindled. Despite this, some fans remain steadfast in their support of the league, religiously attending matches and lending their support to their favourite teams in-person and online. As such, this paper will analyse the reasons behind their lasting passion for the SPL, with an overarching aim of understanding how certain members of dying fandoms around the world remain passionate and unaverred.
16 male SPL fans from a wide range of ages were interviewed, and our research revealed that factors such as habitus, the ability to provide an imagined community in light of wider societal issues, and nationalism played a big role in fandom participation. Furthermore, our study also looked at why the fandom was male-dominated, and how that played an additional role in its appeal.Bachelor's degre
Comorbid cerebrovascular and neurodegenerative burden in mild behavioral impairment and their impact on clinical trajectory
No full textMild behavioral impairment (MBI) is a neurobehavioral prodrome to dementia with multiple phenotypic characteristics. To investigate the complex neurobiological substrate underlying MBI, we evaluated its association with a composite magnetic resonance imaging (MRI)-based measure of concomitant cerebrovascular disease (CeVD) and neurodegeneration; and the interaction effects of MBI and MRI scores on cognitive and clinical trajectory.National Medical Research Council (NMRC)Submitted/Accepted versionThis study is supported by the Singapore National Medical Research Council (grants NMRC/CG/NUHS/2010 and NMRC/CG/013/2013)