INTI Institutional Repository
Not a member yet
2178 research outputs found
Sort by
Comparison of Logistic Regression, Random Forest, SVM, KNN Algorithm for Water Quality Classification Based on Contaminant Parameters
No full textThis study compares four machine learning algorithms Logistic Regression, Random
Forest, Support Vector Machine (SVM), and K-Nearest Neighbors (KNN) in water quality
classification based on contaminant parameters. The purpose of this study is to evaluate and
compare the performance of these algorithms in terms of accuracy. The methodology used includes
data collection, preprocessing, and algorithm implementation with evaluation using crossvalidation
techniques. The results showed that the application of the Stacking method with
Gradient Boosting Meta-learner produced the highest accuracy of 96.00%, outperforming all other
algorithms. In comparison, Random Forest achieved 95.75% accuracy, followed by SVM with
93.25% accuracy, and Logistic Regression and KNN each achieved 90.19% accuracy. This finding
emphasizes that Stacking with Gradient Boosting provides much better performance in water
quality classification compared to other models. This research provides new insights into the
application of machine learning algorithms for water quality management as well as guidance for
optimal algorithm selection
Analysing The Need for Innovative Staff Attendance Tracking System and Designing a Solution for Institut Bakti Nusantara
No full textAttendance at an educational institution or company is important in measuring the performance
and attendance of lecturers. Manual attendance at the Institut Bakti Nusantara (IBN) is still a major
problem, such as time time-consuming and prone to human error. Therefore, this study aims to
analyse the need for an innovative attendance tracking system and design a solution using a QR
Code application for the attendance of lecturers at the IBN. Using the waterfall method which
includes the stages of needs analysis, system design, application development using QR Code
technology, and application testing to test system functionality. The results of this study indicate
that the QR Code Presence application developed can run well according to its function such as
recording the attendance of Lecturers and providing attendance reports. This study produces a
product, the IBN Lecturer attendance QR Code web application that can record attendance and
manage attendance data
Predicting Parkinson’s Disease Using Machine Learning Model
No full textThis research work discusses the steps involved in developing a machine learning program
for the early detection of Parkinson's disease (PD) using a variety of clinical and behavioral
data. By utilizing highlights extracted from persistent data, including engine and non-motor
side effects, the demonstration employs administered learning procedures to identify
patterns indicative of Parkinson's disease (PD). We assess the performance of various
calculations, including back vector machines and neural systems, to determine the most
effective method for accurate forecasts. The results demonstrate the model's potential to
enhance early diagnosis and personalized treatment strategies for Parkinson's infection.
Parkinson's disease (PD) is a dynamic neurodegenerative disorder characterized by engine
side effects such as tremors, inflexibility, and bradykinesia, as well as non-motor side effects
including cognitive disability and autonomic brokenness. Early and precise diagnosis is
essential for effective management and treatment of the infection. In later years, machine
learning (ML) has risen as an effective device in the field of therapeutic diagnostics,
advertising potential changes in the early location and observation of Parkinson's malady
Study on the Relationship Between the Interface Design and the User’s Consumption Level of TikTok App
No full textThe habits of media consumers have undergone a profound shift, with short video content
becoming a dominant trend, particularly among millennials and Generation Z. TikTok, as a leader
in this space, has gained global popularity, yet limited research exists about its consumption
patterns in Malaysia. This study investigates the relationship between TikTok’s interface design
and user consumption levels among Malaysian centennial users (ages 18-24). Using a quantitative
survey of 40 respondents, the findings reveal a high level of TikTok consumption in terms of time
spent and engagement. Statistical analysis via SPSS confirms that TikTok’s interface design
significantly influences user consumption behavior. These findings highlight the critical role of
interface design in shaping the consumption dynamics of short-video apps, and ideally align to
supporting Sustainable Development Goals 12 (SDG 12)
Navigating the Complexities of ESG Integration: Challenges, Opportunities and Path to Sustainable Corporate Development
No full textThis study focuses on the practical application of the Environmental, Social, and Governance
(ESG) framework and examines its role in promoting corporate sustainability and fulfilling
social responsibilities. Through an analysis of theoretical frameworks, literature, and case
studies, the paper highlights the major challenges ESG frameworks face in corporate operations,
including stakeholder conflicts, short-term financial pressures, policy resistance, and
inconsistent quantitative standards. Despite the significant achievements of ESG frameworks
in enhancing brand reputation, attracting investors, and guiding sustainable development, the
research also identifies issues such as "greenwashing," which undermine their overall
credibility and effectiveness. The study suggests that companies must deeply integrate ESG
principles with the United Nations Sustainable Development Goals (SDGs), leveraging
innovation and improving transparency to transform social responsibility into competitive
advantages, thereby driving the global economy's green transition and sustainable development
Exploring Text Recognition Segmentation and Detection in Natural Scene Images
No full textIdentification, segmentation, and recognition of fonts from real-world images are major
challenges in computer vision, particularly due to subtle differences in font shapes, lighting,
and backgrounds. This paper aims to provide a comprehensive review of the latest algorithms
for text detection, segmentation, and recognition from natural scene images. A variety of
techniques are assessed for their use in natural settings, including deep learning-based methods,
region proposal, and feature-based detection. There is additional discussion of the difficulties
of managing changes in text properties such as font type, size, orientation, and noise and
occlusion disruptions. This survey also looks at preprocessing techniques like filtering and
illumination normalization that are meant to increase the accuracy of text detection. In light of
the findings of the literature analysis, this study concludes that the combination of adaptive
segmentation techniques with deep learning-based recognition models offers promising
performance in text recognition in natural scenery images. This survey provides a foundation
for the development of more effective and robust methods for future applications in the fields
of image processing and computer vision
A Comparative Study Between Wireshark and Paessler Router Traffic Grapher (PRTG) in Network Monitoring and Analysis
No full textNetwork traffic analysis (NTA) plays a major role for monitoring network performance, improving
security, and guaranteeing operational efficiency across sectors. This study compares two popular
network analysis tools, Wireshark and Paessler Router Traffic Grapher (PRTG), that handle
distinct areas of network monitoring. Wireshark focusses on packet-level analysis, making it ideal
for comprehensive troubleshooting and security protocol analysis. PRTG, on the other hand,
provides a comprehensive picture, monitoring a wide range of network devices and allowing for
scalability via distributed monitoring, making it ideal for organizations with complicated
infrastructure. This study assesses each tool's capability using essential criteria such as scalability,
real-time monitoring, data management, and security features. By understanding the strengths and
limitations of Wireshark and PRTG, this paper aims to assist network professionals in choosing
the most suitable tool for their specific monitoring and diagnostic requirements. Practical
recommendations are provided to guide both beginners and experienced users in leveraging these
tools effectively
A Study On AI-Driven Solutions for Cloud Security Platform
No full textCloud computing has developed as a reliable approach to adopting various services inherent in
data management but has its weaknesses in terms of security risks such as unauthorized access,
data leakage or any other threats from insiders. This paper examines the role of AI in the
improvement of cloud security with specific emphasis on deep learning, ensemble learning and
lightweight AI approaches. Cognitive tasks comprise integration, computational cost, and the
ethical effect of the algorithm are identified and discussed. Real-world applications and
possibilities for further development, such as federated learning and XAI, are also described in
order to give recommendations for the effective application of AI-based cloud security. Finally,
this research seeks to help organizations protect cloud structures and resources using intentioned
AI solutions
Enhancing Creative Content in Education Delivery Utilising Multimedia Applications
No full textThe shift to online education due to the Covid-19 pandemic has transformed the delivery of
educational content, a practice that continues in most higher education institutions. This study
investigates how multimedia applications can improve the way innovative information is delivered
in the classroom. To get information about their online learning experiences, 120 university
students participated in online surveys and interviews using a quantitative methods approach. The
study investigates the difficulties students encounter in reaching successful learning objectives and
how these difficulties relate to the information learned through online means, as opposed to inperson
lectures. Results show that students' focus and participation in online classrooms are greatly
impacted by technological obstacles, such as subpar gadgets and erratic internet connections.
These observations highlight the necessity of utilising multimedia applications and addressing
technology constraints in order to enhance the educational process and close the gap between
traditional classroom settings and online learning environments. This research outcome strongly
addresses the targets within Sustainable Development Goals 12 (SDG 12)
Application of Artificial Intelligence in Healthcare Industry: A Critical Review
No full textLong-term cost-effectiveness studies are needed to determine the financial impact of AI
adoption on healthcare systems. Upskilling the healthcare workforce will be vital to ensure
professionals can adapt to evolving AI-driven workflows. Developing standardized
frameworks will be crucial for seamlessly integrating AI solutions across different healthcare
institutions. Beyond the immediate challenges these journals identified, other practical
considerations deserve attention. AI in healthcare presents a powerful opportunity for
transformation. However, acknowledging and addressing the ethical, practical, and logistical
challenges can pave the way for responsible development and ensure AI fulfils its
transformative potential, ultimately improving healthcare for all. In conclusion, while AI holds
immense promise for the future of healthcare, its successful integration hinges on addressing
these critical issues. To translate these findings into practical steps, a multi-pronged approach
is necessary. Further research on the effectiveness of AI in various settings and clear regulations
are necessary to ensure AI is implemented fairly, ethically, and effectively across the globe