Journal of Advanced Research Design

Development of Mobile Application for Enhancing Emotion Management for UTHM Students using DASS-21

2025-02-04T10:34:34+07:00

Managing one's emotions is a crucial skill, particularly for university students facing intense academic and social pressures in today's competitive environment. However, Malaysian university students often struggle with limited emotional awareness, which further compounds their mental health challenges. Therefore, a mobile application called Therapist was proposed to offer assistance for university students in managing their emotion based on DASS-21 inventory model. The mobile application was developed using a Multimedia Mobile Content Development method. Usability testing was conducted with the target user group using the System Usability Scale. A score of 86 was obtained from the results, which is "acceptable" in the Acceptability Ranges scale's. This corresponds to a "B" grade on the Grade Scale and an "Excellent" Adjective Rating. The results suggest that the Therapist application can be a valuable tool for enhancing students' emotional management capabilities especially for Universiti Tun Hussein Onn Malaysia (UTHM) students.

Implementing Multiple Dynamic Flip Algorithm: Bread on Mobile Case Study

2025-03-03T14:26:30+07:00

Mobile apps developers encounter a range of challenges and problems during the development. These issues can vary depending on the platform (iOS, Android, cross-platform) and the specific project, but here are some common problems faced by mobile app developers. The need for a structured mechanism in mobile apps development is urgently crucial to reduce the cost of manpower charges in conjunction with the simple and easy framework to assist mobile apps developer. In response to these scenarios, this paper introduces the Multiple Dynamic Flip Algorithm (MDFA), a new algorithm in developing mobile-based applications that serves as the mobile app’s framework. The algorithm developed allows the data to be stored in the database and the administration of the system can only change the data according to the categories of data that have been systematically arranged in the database. This technique allows changes to information and data to be done automatically without requiring users to run back to mobile-based programming. Process of updating information needs to be generated back in mobile-based programming and needs to be uploaded back in the play store and users must download the application software back. By using this MDFA algorithm, the problem is no longer needed and can be solved more quickly. Implementation results of MDFA via university-industry real case study through Bread on Mobile (BOM) apps used for mobile apps bread ordering.

Adaptive Hybrid Reduced Differential Transform Method in Solving Nonlinear Schrodinger Equations

2025-03-24T14:59:10+07:00

In this paper, piecewise-analytical and numerical solutions are obtained by a new adapted approach named the Adaptive Hybrid Reduced Differential Transform Method (AHRDTM). The fundamentals of the method are introduced, followed by its application to Nonlinear Schrodinger Equations (NLSEs). Analytical and numerical solutions are acquired using piecewise convergent series with computationally feasible components across a sequence of sub-intervals of varying length. This succeeds due to the adaptive algorithm introduced and the substitution of the non-linear term in the NLSEs with their corresponding Adomian polynomials. The accuracy of the AHRDTM is observed through numerical comparisons between the proposed method and the Modified Reduced Differential Transform Method (MRDTM) with their respective exact solutions. The absolute errors presented in the table reveal that the proposed approach has better accuracy in solving the considered equations. The results and pictorial illustrations have been provided to demonstrate the reliability of the method’s accuracy in obtaining approximate solutions.

Uncovering Depression on Social Media using BERT Model

2025-03-17T13:36:56+07:00

The lack of early detection and effective treatment programs for depression has left millions struggling with mental illnesses, including anxiety, sleep disorders and, in severe cases, self-harm or suicide. Adolescents and young adults, who are among the most active users of social media platforms like Twitter, represent a particularly vulnerable demographic. With Twitter often serving as a digital diary where individuals share thoughts and emotions, its data offers a unique opportunity for mental health research. This study explores the potential of leveraging Twitter data to detect depression and identify at-risk individuals. Using advanced machine learning techniques, including Natural Language Processing (NLP) with the BERT (Bidirectional Encoder Representations from Transformers) model, this research builds a predictive system that analyses linguistic cues, sentiment and user interactions. The methodology involves integrating and preprocessing data from multiple sources, ensuring a comprehensive and reliable approach to model development. The findings demonstrate the BERT model’s superior accuracy, outperforming traditional machine learning methods like Logistic Regression and Support Vector Machines. This study underscores the effectiveness of social media as a tool for early detection and intervention, offering an innovative approach to mitigating the global burden of mental health disorders.

Regularized Stacked Autoencoder with Dropout-Layer to Overcome Overfitting in Numerical High-Dimensional Sparse Data

2025-03-22T08:16:04+07:00

High-dimensional sparse numerical data are normally encountered in machine learning, recommender systems, finance and medical imaging. The problem with this type of data is that it has high dimensions (many features) and highly sparse (most values are zero), which is prone to overfitting. The data visualization can be achieved through a neural network architecture called stacked autoencoders. These multilayer autoencoders are designed to reconstruct input data, but overfitting is a major problem. To overcome this problem novel L1 Regularization-dropout technique is introduced to reduce overfitting and boost stacked autoencoder performance. L1 regularization penalizes large weights, simplifying data representations whereas the dropout technique randomly turns off neurons during training and makes the model dependent only on the selected turn-on neurons. The model employs batch normalization to improve the performance of the autoencoder. The approach was implemented on a high-dimensional sparse numerical dataset in the field of cybersecurity to minimize the loss function, measured by Mean Square Error (MSE) and Mean Absolute Error (MAE). The findings were compared to the conventional stacked autoencoder. The study revealed that the suggested method effectively mitigated the issue of overfitting. Stacked autoencoders, when combined with L1 regularisation and the dropout approach, are very successful in handling high-dimensional sparse numerical data in a diverse range of applications.

A Comparison Study on User Acceptance of Digital Healthcare Services among M40 and B40 Community in Selangor: A Study on SELangkah Application

2025-03-06T12:12:15+07:00

Over time, the global digital healthcare industry has started to take shape with much effort has been put into improving the effectiveness and accessibility of the healthcare system in Malaysia for the benefit of its people. The Selangor state government is working to make all communities safer and healthier places and one of its initiatives is the SELangkah application. Nonetheless, there are still concern over the take up of digital healthcare services among citizens especially those in the low-income and middle-income communities; B40 and M40. Previous study has shown that different economic status has an impact towards the acceptance level in technology services. Therefore, this study utilises qualitative research method to investigate the state of user acceptance in digital healthcare services among B40 and M40 users in Selangor through the case study of SELangkah application. Focus group discussion was conducted with over 10 participants from Selangor to evaluate and compare the acceptance level of digital healthcare services between both income group with the instrument development were guided by the Extended Technology Acceptance Model (ETAM). The area of group discussion covers users’ perceived usefulness, perceived ease of use, attitude, behaviour, external challenges and actual use. The findings show that both M40 and B40 group find the application to be beneficial; in addition, they believe there should be more awareness campaign to promote the application to the public. A few clear differences between the acceptance level of the two groups would be the simplicity of the language used in the application, the recommended features and future needs between the two groups. M40 group find no difficulties to navigate the application but B40 group mention that the used words can be simplify. While M40 group recommended for gamification and high engagement features; the B40 group is looking more at practicality such as more information or linkages to nearby clinic. Nevertheless, both groups did not find major difficulties and they see a lot of benefit of the application. This will form the fundamental basis for future research efforts aimed at closing the technology acceptance gap.

Hybrid Multilayer Perceptron Neural Network for Transformer Health Index Monitoring

2025-03-03T12:19:37+07:00

Power transformers are critical to electrical systems, requiring accurate health monitoring to ensure reliability and prevent failures. Traditional assessment methods often fail to capture complex variable interactions, resulting in suboptimal maintenance strategies. This study introduces a hybrid multilayer perceptron (HMLP) neural network for transformer health index (HI) monitoring, using approximately 500 data points from the Klang Valley, including gas formation data. The HMLP is benchmarked against classifiers such as multilayer perceptron (MLP), K-Nearest Neighbors (KNN), Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM), outperforming them with 92.35% accuracy and a mean squared error (MSE) of 0.78. Additionally, three training algorithms; Backpropagation (BP), Levenberg-Marquardt (LM) and Bayesian Regularization (BR) were tested, with the BR algorithm achieving the best performance at 94.13% accuracy and an MSE of 0.39. This research highlights the potential of the HMLP network, particularly when trained with BR, to revolutionize transformer maintenance by enabling precise THI predictions, facilitating proactive interventions and ensuring power system reliability.

Designing Enhanced Automatic Genetic Clustering Algorithm for Unknown Number of Clusters: An Experimental Evaluation

2025-01-21T15:14:10+07:00

This research seeks to develop an improved automatic genetic clustering technique to support the negotiation of heterogeneous and multi-attribute cloud services. It is known that heterogeneous cloud services involve cloud service providers offering varying sizes and costs for cloud service virtual machines, processors and storage. Existing automatic clustering techniques, such as the Automatic Genetic Clustering Algorithm for Unknown Number of Clusters (GCUK) algorithm, can provide the solution for heterogeneous cloud services negotiation, but they can also result in sub-optimal clusters, overlapping and partial solutions, local optima trapping and sub-optimal chromosome size. The automatic clustering technique employs the modified GCUK algorithm to address this issue. It was intended to solve GCUK constraints, local optima, imperfect clustering and suboptimal chromosomal size, as well as to support different cloud services. The findings indicate that the improved GCUK algorithm can provide optimal clustering solutions and avoid local optima, achieve optimal accuracy and resilience and obtain dynamic and optimal chromosomal sizes at runtime.

The Assessment of Tall Building Structure Performance Due to Seismic Effect in Malaysia

2025-02-14T09:47:04+07:00

Recent natural disasters have brought attention to the vulnerability of buildings in Malaysia, resulting in severe damage, loss of life, and economic setbacks. Many existing reinforced concrete structures in the country were not adequately designed to withstand seismic forces, as they primarily focused on vertical loads. Therefore, it is crucial to evaluate the seismic performance of buildings to ensure their safety during earthquakes. This research aims to develop fragility curves specific to tall concrete buildings in Malaysia. The study analyzes two building models, differing in the number of stories: a 30-story type I building and a 25-story type II building. The analysis involves pushover analysis and incremental dynamic analysis, considering gravity and wind loads as per the Malaysian standard practice (BS8110). Lateral design loads are utilized for the pushover analysis, while five records of far-field ground motion are employed for the incremental dynamic analysis. Furthermore, different soil types are taken into consideration to assess their impact on the structural response. Fragility curves are produced through the examination of inter-story drift ratios recorded during the evaluation of five distinct ground motions for each soil classification, along with incremental variations in peak ground acceleration. The structural performance is evaluated using five distinct performance levels: operational phase (OP), immediate occupancy (IO), damage control (DC), life safety (LS), and collapse prevention (CP). Fragility curves are developed for both building models, considering the four soil types. The findings reveal that the inter-story drift ratio or damage measure values increase with higher peak ground accelerations. Additionally, the pushover analysis indicates that the 30-story model (type I building) exhibits higher base shear compared to the 25-story model (type II building), suggesting a correlation between base shear and building height. A total of eight fragility curves are constructed, representing the probability of exceeding IO and CP limit states. Notably, under soil type A bedrock with a peak ground acceleration of 0.1 g, the 30-story building demonstrates better performance compared to the 25-story building, with maximum probabilities of exceeding CP damage states of 4% and 8% respectively. The fragility curves highlight the significant influence of soil conditions on structural vulnerability.

Effective Periodic Noise Reduction for Ship Corrosion Image

2025-03-19T10:37:24+07:00

Marine corrosion research has recently applied advanced image processing techniques. However, the investigation might have been interrupted due to the frequent disruption of corrosion images by certain noises, leading to a significant reduction in analysis accuracy. The comprehensive investigation of corrosion images required the implementation of progressive noise reduction techniques, especially for addressing periodic noise. Hence, the primary objective of this research is to investigate the reduction of periodic noise in ship corrosion images by leveraging a state-of-the-art image processing technique known as the Modified Gaussian Notch-Reject Filter (MGNRF). The methodology includes the selection of noise frequency using a window-based approach and the application of MGNRF for noise reduction. In comparison to other algorithms, the proposed algorithm has achieved promising results, with a mean absolute error (MAE) of 1.287, a remarkable average peak signal-to-noise ratio (PSNR) value of 40.215 and an exceptionally high mean structural similarity index (MSSIM) value of 0.998. This demonstrates the efficiency of the proposed algorithm in efficiently reducing periodic noise and enhancing the quality of ship corrosion images, thus ensuring accurate analysis and making a significant contribution to this critical domain.

Accurate Corrosion Detection and Segmentation on Ship Hull with Pixel Property Method

2025-03-19T10:42:52+07:00

The ship's hull is primarily exposed to salt-laden sea spray and high moisture, making it susceptible to corrosion. This has become a major issue in the shipping industry, as corrosion weakens the ship's structural integrity, necessitating expensive maintenance and posing safety concerns. Despite the latest advancements in corrosion maintenance technology, it is essential to detect corrosion as early as possible using computer vision or image processing techniques. However, both approaches have limitations when it comes to detecting weak corrosion boundary and blurry prominent corrosion features. Therefore, the primary objective of this research is to accurately detect corrosion boundaries on the ship's hull using pixel property method. Firstly, data acquisition is performed to identify suspected corrosion regions on the ship's hull. Next, a threshold is calculated by averaging 100 corrosion images of the ship's hull. Afterward, each pixel in the image is analysed to determine the connected components of the corrosion areas. The pixel list and area coordinates are collected after analysing all connected components. Large connected components are merged into a single larger region using morphological closing and flood-fill operations. Finally, the pixel property method is applied using the pixel list and area coordinates to accurately detect corrosion boundaries on the ship's hull. According to the results, the proposed method successfully detected corrosion regions on the ship's hull with a high level of accuracy. Furthermore, the robustness of this method was demonstrated by its ability to segment the weak corrosion boundary and blurry prominent corrosion features on the ship’s hull. These findings indicate that the proposed method is highly accurate for detecting corrosion on ship hulls.

Comprehensive Review on Sensitivity Determination in Fiber Bragg Grating Accelerometer

2025-01-15T14:11:02+07:00

Achieving accurate sensitivity measurements in Fibre Bragg Grating (FBG) sensors is crucial for their reliable application across various fields. This review paper delves into the sensitivity characteristics of FBG sensors, emphasizing the importance of precision in sensitivity determination. Various models and approaches for FBG strain, temperature, pressure, displacement and acceleration sensors have been examined. Among these, FBG acceleration sensors (FBG accelerometers) present the most significant challenge due to the dynamic nature of the measured acceleration in terms of both magnitude and frequency. Studies have shown that sensitivity in these sensors is inherently frequency-dependent. The assumption of linear sensitivity within one-third or one-half of the resonant frequency can be improved through advanced methods, including machine learning techniques. By analysing both theoretical and experimental studies, this review highlights the critical factors affecting sensitivity accuracy and identifies areas where discrepancies may arise. The aim of the review is to provide a thorough understanding of the methods for achieving accurate sensitivity measurements in FBG accelerometers, thereby guiding future research and development in this domain.