UTAR Institutional Repository (Universiti Tunku Abdul Rahman)
Not a member yet
6132 research outputs found
Sort by
Hotel recommendation system using machine learning
Get PDFIn recent times, choosing the appropriate hotel destination and making bookings has become
increasingly complex due to the rapidly growing volume of available online information. The
importance of recommender systems (RSs) is rising as they help users make informed decisions
and provide comprehensive insights into products or services. Managing user-generated data such
as votes, ratings, views, and reviews presents significant challenges. There are three objectives in
the study, which is to perform data preprocessing on the Google Reviews dataset for hotels in
Perak using an instant data scraper, to develop three suitable machine learning models on the
cleaned dataset and evaluate their performance, and to propose a recommendation system based
on the developed machine learning models. The methodology includes data scraping,
preprocessing, implementation of machine learning techniques such as Naïve Bayes, Random
Forest, and Support Vector Machine (SVM), and proposes a recommendation system. The system
integrates these models to provide hotel recommendations based on each user's preferences. The
results show that the proposed model is effective and generates recommendations for the user. The
future work includes expanding the dataset, refining the recommendation algorithm, using natural
language processing techniques with the addition of multilingual reviews, and deploying the
system as a user-friendly application or mobile application
Intelligent medicine box system with AI-powered pill detection and IOT integration
Get PDFTaking medication as prescribed is a challenging task for most patients, particularly for those with busy lifestyles. In simple words, medication adherence can be defined as taking prescriptions at the right time and in the correct dosage. Adhering to medication schedules and dosages is crucial for managing chronic conditions such as hypertension and high cholesterol. Failure to follow the medication regimen can lead to several adverse consequences, including disease progression and the deterioration of health conditions. In the long run, this can ultimately reduce the overall quality of life, leading to an increased risk of long-term health consequences. To address this issue, an Intelligent Medicine Box System that leverages the power of Artificial Intelligence (AI) and the Internet of Things (IoT) has been developed to improve the user’s medication adherence. This book describes the development of an Intelligent Medicine Box System with AI-Powered Pill Detection and IoT Integration. The Intelligent Medicine Box System, which leverages IoT, provides the features of timely reminders via mobile application and the buzzer in the pill box to ensure the user takes their medication as prescribed. Besides that, the risk of running out of medicine will be reduced with the real-time pill tracker feature that automatically tracks and monitors supply levels and alerts users when refills are needed. In addition, the AI-powered pill detection and counting feature using a deep learning model can further detect foreign objects and potential missed doses, thus reducing contamination and enhancing medication adherence. By ensuring their medications are taken on time, this system can help in managing users' conditions more effectively, thereby improving their quality of life
Web-based clinic management system
Get PDFThis project focus on the development of web-based clinic management system. The system which are currently used by others clinic, contain a few problems and limitations. Nowadays, the clinic management system may contain some lack of feature such as BMI calculation and record storing system, an effective clinic management system and a well-structured and managed billing and health report system. Therefore, this project and development of system is carried out to resolve the limitation and problem of the clinic management system others clinic currently used. In this project, a real-time appointment system has been carried out to allow patient to make their appointment faster and get confirmation immediately. Additional tools of the clinic management system website also have been carried out such as BMI measuring tools which can always track the user BMI result, make comparison and give the appropriate feedback. Also, the website allow patient to pay their clinic bill online and they can choose their preferred method to pay their bill and ensure safety to pay through online. On the other hand, the system included using incremental methodology to divide the project accordingly. There are a few technologies and tools will be used in this project to carry out a well-structured and managed clinic management system
Multidimensional poverty and anti-poverty strategies in urban China: A mixed-methods study of Shandong province
Get PDFOver the past few decades, China has made remarkable strides in its battle against poverty. However, these efforts have been primarily concentrated in rural areas, while urban poverty has been neglected to some extent. China has not established specific poverty standards for urban areas, nor has it implemented unified measures for urban poverty as it has for rural areas. This research aims to comprehensively explore the overarching landscape and key contributing factors of multidimensional urban poverty in Shandong Province and investigate the lived experiences of its poor households, in order to provide recommendations for urban anti-poverty strategies in the region. This research employs a mixed-methods approach. The quantitative research utilizes a Dual Cutoff method to examine the overarching landscape of urban poverty in Shandong Province and applies logistic regression analysis to identify its contributing factors. Building upon the quantitative findings and relevant literature, the qualitative research employs an interpretative phenomenological approach to investigate the challenges faced by urban poor
households in Shandong Province, as well as their coping strategies and poverty alleviation needs. The quantitative and qualitative research findings were then compared and discussed.
The quantitative data analysis reveals that the overall reduction in urban poverty in Shandong Province over the past few years is primarily due to the decrease in the incidence of poverty rather than the intensity of poverty. Logistic regression analysis identified low educational attainment, chronic disease, large household size, gender, and poor surrounding environment as the primary factors contributing to urban poverty. Through qualitative data analysis, this research unveiled that urban poor households mainly faced challenges related to financial hardship, health dilemmas, poor housing conditions, and ineffective government administration. The qualitative findings supplemented the quantitative results by emphasizing the role of government administration in urban poverty. By integrating the findings of quantitative and qualitative research, this study offers targeted recommendations for effectively addressing urban poverty in the region
Mobile application for detecting autism spectrum disorder (ASD)
Get PDFAutism Spectrum Disorder (ASD) is a neurodevelopmental disability that affects how humans interact, communicate, and behave. Autistic people often find it hard to socialise with others and may engage in self-injurious behaviours. Specifically, there is no cure for this disorder. Additionally, it is expensive to detect ASD as it requires long-term monitoring by experts. Some diagnostic methods even involve brain scanning. Thus, it is unaffordable for most families, especially those with limited financial resources, even if their children suffer from this disorder.
Early intervention is important, making early detection of ASD particularly significant. To address this challenge, face recognition technology powered by deep learning has emerged as a promising diagnostic tool. This project aims to implement transfer learning using facial recognition to detect ASD. While researchers have proven that questionnaires can be effective screening tools with good detection accuracy, this project seeks to further validate the accuracy of the detection process. The project provides two separate methods, which are facial detection and Q-Chat 10 approaches to streamline the ASD detection process
Enhanced removal of pharmaceuticals and personal care products (PPCPs) using sunlight-driven photocatalyst (g-C3N4/BiNbO4)
Get PDFOxytetracycline (OTC) is increasingly recognized as an emerging contaminant among pharmaceuticals and personal care products (PPCPs), posing significant risks to both human health and the environment by potentially disrupting the central nervous and digestive systems. Conventional treatment methods have proven inadequate for degrading OTC due to its unique chemical properties. This study investigates the degradation of OTC using a g-C₃N₄/BiNbO4 composite photocatalyst. The photocatalyst was comprehensively characterized by FESEM-EDX, XRD, FTIR, and UV-Vis DRS to examine its morphology, elemental composition, microstructure, crystal structure, functional groups, and bandgap energy. FTIR analysis confirmed the presence of functional groups such as O–H, C–N, and C=N within the composite. UVVis DRS analysis determined the bandgap energies of pure BiNbO4 and the optimized g-C₃N₄/BiNbO4 composite to be 2.87 eV and 2.85 eV, respectively. Moreover, photocatalytic experiments revealed that the composite containing 0.3g-C₃N₄/BiNbO4 achieved an impressive removal efficiency of 98% within 180 minutes. These findings underscore the potential of the g-C₃N₄/BiNbO4 photocatalyst as an environmentally friendly solution for degrading various organic pollutants under sunlight irradiation.
Keywords: Oxytetracycline, emerging contaminants, g-C₃N₄/BiNbO4 composite, UV–Vis diffuse reflectance spectroscopy (UV-Vis DRS), bandgap energy Subject area: QD701-731 Photochemistr
Reducing material losses in a blending production line - A case study in an air filter manufacturing company
Get PDFMaterial losses are a common issue in production lines at manufacturing companies. Even with advanced technology invested in the production line, material losses still occurred due to various factors, such as material handling and transportation. This study was conducted at Factory X, an air filter manufacturing company located in Perak, Malaysia, with the main objective of reducing material losses in the chemical blending production line by 5%. The DMAIC (Define, Measure, Analyse, Improve, Control) approach was chosen as the primary methodology, supported by on-site examinations, Material Flow Analysis (MFA), the 5 Whys analysis, and Kaizen. A framework was developed to guide the step-by-step implementation in the factory, keeping the project on track and ensuring a systematic improvement process. After measuring and analysing data before and after the test run, the results showed a reduction in material losses, decreasing from 17.86% to 11.97%, representing an improvement of 5.89%. The study demonstrated the success of the framework and proposed several recommendations for future research, including testing the framework in different production lines and industries, applying it to other sections of the same production line, optimising MFA for more accurate measurement, and investigating alternative materials for reducing material losses.
Keywords: Blending Production Line; Material Losses; Manufacturing System; Material Flow Analysis (MFA); FESEM Analysis Subject Area: TS155-19
Application development for traditional chinese medicine herb
Get PDFTraditional Chinese Medicine (TCM) has been practiced for thousands of years, relying heavily on the accurate identification and combination of herbs for effective treatment. However, the vast array of herbs and their subtle differences pose significant challenges in ensuring accurate identification and proper use, especially for non-experts. To address these challenges, this project proposes the development of a mobile application that utilizes advanced image recognition technology, specifically Convolutional Neural Networks (CNNs), to accurately identify TCM herbs and provide users with personalized herb combination recommendations.
The application is designed to enhance the accessibility and safety of TCM practices by offering a user-friendly interface that supports herb identification, detailed information on each herb, and symptom-based herb recommendations. The project leverages the capabilities of a pre-trained CNN model, fine-tuned with a comprehensive herb dataset, to deliver high accuracy in herb recognition. Additionally, the application includes features such as user authentication, camera integration for herb scanning, and a history tracking module to support personalized user experiences.
Through this project, the application aims to modernize TCM practices by integrating cutting-edge technology, providing a valuable tool for practitioners, students, and patients to enhance their understanding and effective use of TCM herbs
Impact of training, compensation, and workload on employee job performance in SMEs in Malaysia
Get PDFThe purpose of conducting this research is to investigate the factors that are affecting employee job performance in SMEs in Malaysia. The factors that may have significant impacts on job performance are training, compensation, and workload. The research was conducted by distributing questionnaires by using Google forms to the target respondents, which are employees in SMEs in Malaysia. Thus, the questionnaire was distributed through Google Forms, and the researchers successfully collected 384 responses. Statistical Package for the Social Sciences (SPSS) has been used to analyse and interpret the data collected for pilot study and pilot study. Additionally, our study employed multiple regression analysis to examine the significance of the impact between the independent variables (workload, training, and compensation) and the dependent variable (employee job performance). In this research, the three independent variables (training, compensation, and workload) are examined that there has a significant impact on the dependent variables (employee job performance). Thus, the detailed results of the research, limitation, and recommendation will be further discussed in the chapters below.
Keywords: Employee Job Performance; Training; Compensation; Workload; Small and Medium-Sized Enterprises Subject area: HF5549- 5549.5 Personnel management. Employment Managemen
Perak tourism search using Google map Android app development
Get PDFBachelor of Computer Science (Honours)
Faculty of Information and Communication Technology (Kampar Campus), UTAR
iv
ABSTRACT
In the age of globalization, the Internet plays a vital role in people’s lives. To stay connected to the Internet, various technologies, including mobile devices have undergone rapid development. Therefore, various mobile applications are developed to fulfil the growth of user’ demand in various sectors, such as the tourism sector. Hence, this project will delve into the functionalities and features proposed by the existing systems and their limitations. Based on the analysis on the current applications, this project aims to develop a tourism search application to overcome the limitations by proposing innovative solutions. It seeks to enhance user experience by addressing issues like unnecessary features that burden the system, imbalance food priority and fake reviews. It will focus exclusively on tourism search to provide a simple, straightforward user interface for the user by eliminating non-essential features.It will include a categorization system for easy exploration and discovery of various cuisines options. Real identity verification via face detection and face recognition is implement to maintain trust and authenticity on our platform, therefore minimize the fake review. This project implements Agile Development for the whole development processes. This application is developed using Flutter with Dart programming language. Google Map API is integrated to display the route, calculate duration and total distance between user’s location and destination. Firebase cloud-based data management server is integrated and its services, such as authentication and firestore are utilized. Supabase storage is integrated to manage the photos