8333 research outputs found
Sort by
Heart Disease Prediction: Technical Report
No full textThis report details the development of a heart disease prediction model using machine learning techniques. The primary objective was to create an accurate and reliable model capable of predicting the presence or absence of heart disease based on individual health attributes. The dataset, comprising demographic and health-related information, was explored and visualized to gain insights into the underlying patterns.
The Random Forest classification model was selected and optimized through grid search for hyperparameter tuning. The model achieved a commendable accuracy on the test set, demonstrating its effectiveness in predicting heart disease. The exploration of key features and their impact on predictions was crucial in understanding the model's decision-making process.
The project's findings reveal notable trends, such as age distribution and target variable proportions. The Random Forest model, with its tuned parameters, outperformed baseline models, showcasing its potential for real-world applications. The classification report and confusion matrix provide a detailed assessment of the model's strengths and areas for improvement.
In conclusion, this project successfully developed a machine learning model for heart disease prediction, with the Random Forest algorithm emerging as a robust choice. The report recommends further exploration of alternative models, feature engineering, and the inclusion of diverse datasets to enhance predictive capabilities. The deployed Random Forest model, saved for future use, holds promise for aiding medical professionals in early detection and intervention for individuals at risk of heart disease. The findings of this report contribute to the ongoing efforts to leverage machine learning in healthcare for proactive and personalized patient care
URVote: Technical Report
No full textThe main purpose of this technical report is to express and explain the development of my Computing project, which URVote web application for the midpoint, the URVote project is to be completed for my four-year computing project. This project aims to use a new technique of casting a vote, the URVote is mostly developed for organisations to get employees to vote for any new policy regulation or issue. This report defines the 7 implementation, and functional requirements of URVote developed so far example use case diagrams, and other important web applications, the admin, employer and managers prefer such web applications for their companies to get the ideas of his/her employees staff worker through an easy web App or any technology that can be used and gather information as quickly as possible which that can prosperity support them. Furthermore, I will add additional information to this technical before the final submission. URVote web app will allow users such as staff to vote yes or no, one voter can only post one vote on the argument, and at the end, the admin will receive a feedback report which will benefit the admin in many ways to know what the employee is seeking for, which I am in opinion that will help the admin a lot
Switchify: Technical Report
No full textThe proposed home automation system which constitutes an ESP32 microcontroller coupled with a Mobile application serves the purpose of giving users a wide range of options that allow them to control or monitor smart devices relevant to their homes from wherever and whenever they want. The Mobile app enables complex managed devices, and it improves Firebase to show actual sync information in real time. The ESP32 is the main controller and manager of the whole system, establishing a safe and reliable Wi-Fi connection for effective communication.
Some main features such as a pairing algorithm based on secure connection, device controlling for lights switch on and other gears as well as the automation of lighting scene modes are involved. Our app provides easy-to-use Flutter interfaces, making it possible for the software to run on both mobile platforms.
The program comprises of Dart programming for the mobile app, Arduino IDE for ESP32 firmware, and Firebase integration for cloud-based storage. The GUI features a menu bar with options that allow the user to set up time schedules, modify scenes, and give reviews.
The forthcoming improvements stand through the introduction of the energy efficient algorithms and thus the data logging for in-depth analysis. The system presents a scalable and secure solution for smart home automation, which increases user convenience and reduces energy consumption
Diversity and Inclusion: A comparative study of IBM (India) and Intel (Ireland)
No full textIn current severely evolving business world, diversity and inclusion are essential components of an effective organisational structure. This dissertation delves into the execution of specific initiatives and programmes, their impact on employee engagement, and the challenge of attaining organisational performance. The qualitative research study offers methods like employee semi-structured interviews. The comparison of diversity and inclusivity initiatives, their differences, how they affect employee engagement, and whether or not they improve organisational performance were the main research questions that needed to be addressed.
Several themes emerged from the analysis of the interview data with the help of theme coding: the choice and execution of effective initiatives and programmes, the impact of these initiatives on the specified performance level and employee engagement, as well as some potential opportunities and challenges for the future. To address the needs of diversity and inclusion, IBM and Intel, for example, have each defined, developed, implemented, and evaluated a number of important initiatives, including employee resource groups, diversity training, and targeted recruitment. As evidenced by the higher levels of commitment, motivation, and belonging, it has a positive effect. They could accomplish the performance at the same time by being innovative, improving their problem-solving skills, and trying to comprehend the demands of a wide range of clients.
However, issues like implicit prejudices, the need to build strong bonds with people from different cultural backgrounds, and the increased resistance to change force organisations to assess the suggested approaches. More cross-cultural training, marketing, effective leadership, and the technological tools that businesses already employ for talent management are among the other suggestions. Planning potential diversity and inclusion policies and identifying the most effective initiatives and programmes could both benefit from such research studies
Examining the Relationship between Transformational Leadership and Intrinsic Motivation among Employees in ECCO Store in Ireland
No full textPurpose: The aim of this study is to explore the connection between transformational leadership and intrinsic motivation among staff at Ecco Store in Ireland to help improve the leadership style.
Methodology: The Research will use both qualitative and quantitative methods. The study was also qualitative, and questionnaires were used to obtain the views of the respondents. Both quantitative and qualitative methods was used, and it is justified to strengthen the validity of the results and spot any discrepancies.
Findings: The study found out that there is a positive correlation between transformational leadership and intrinsic motivation. Also, management considers talent acquisition, and the employees are selfdriven on all activities. Training helps develop employee skills and thus motivates them.
Practical Applications: Effective governance practices will benefit the retail sector's progression and therefore boost occupation in the nation. This is one of the noteworthy worldwide concepts of obliteration of abject poverty. Through the help of human resources, organizations can reform their policy to help the staff benefit from their supervisors intentionally.
Research Limitations: The research was conducted only in Ireland Ecco store; however, Ecco Store have different outlets across the globe. Therefore, the results might not depict the Ecco leadership wholly.
Originality and Value; There is no research that has been done on the motivation of the Ecco store Employee. Therefore, this work is tailored only for the Ecco store. And the results can be significant to other retailers dealing with similar or competitive products
Adversarial Resilience in Malware Detection: A Two-Stage Structural Analysis Approach for Robust Cybersecurity
No full textIn the fast-changing world of cybersecurity, more and better computer virus attacks are becoming a big issue for keeping digital security safe. This paper suggests a strong two-step system for detecting Malware. It combines simple malware detection and advanced harmful software detection to better identify and fight all types of unwanted programs, even when they use clever hidden methods. Using big computer learning techniques like Logistic Regression, Random Forest and Decision Tree plus the deep thinking skills of Long Short-Term Memory (LSTM), this system shows high results in correctly putting harmful software or normal files. The test, done on a big and complete BODMAS data set, shows the system’s strength against harmful attacks from malware creator. It proves it can change quickly to deal with fast-moving malware or bugs in computer systems. The LSTM model stands out, getting it right 82% of the time. This shows that it is very good at spotting hard to find patterns linked with bad behavior. Even though it’s been successful, the system knows that we need a bigger and more varied set of information to make itself even better. The study helps make computer security stronger by suggesting a new and changeable way to spot malware. It shows the need for constant improvement and growth in order to properly fight off changing threats over time
A lightweight failure management pattern for microservices
No full textThe project develops a novel lightweight failure management pattern for distributed transactions in microservices. We focus on issues related to data consistency, fault tolerance, and atomicity. As the adoption of microservices architecture becomes prevalent due to its advantages over monolithic architecture, concerns arise regarding the coordination of distributed transactions and the preservation of data consistency across microservices. The specific scenario explored involves a web application with interconnected microservices, highlighting the potential for inconsistent data and performance issues in the face of microservice failures.
To address these challenges, the research proposes a novel lightweight failure management pattern for microservices. It takes the bare-minimal fault-tolerant features of Two-Phase Commit and Saga Cloud Pattern, which are the retry mechanism and temporary databases. The 2PC protocol ensures atomicity, reliability, and data consistency in distributed transactions and Saga patterns automate the failure recovery functionality. The unique aspect of this solution lies in the incorporation of a temporary database, serving as a safeguard against failures. In the event of a microservice failure, transactions are logged in the temporary database. Upon the restoration of the failed service, it retrieves the data from the temporary database, and the 2PC protocol verifies the availability of all participating services before finalizing the transaction and Saga pattern automates the regaining original state of the microservice.
To demonstrate the expected results, a web application (called My ArtGallery) consisting of two microservices (Admin and Main) was developed. The implementation involves React JS for webpages, Python and Django for backend processing, MySQL as the database, RabbitMQ for message publication and subscription, and Docker for deployment. Two use cases demonstrate the main features of the proposed algorithm with encouraging results. We therefore think the novel lightweight failure management pattern can be of great academic and practical interest
Privacy Impact Assessment of Third-Party Dependencies
No full textThis research focuses on developing a novel framework for Privacy Impact Assessment (PIA) within the context of third-party dependencies in software development. The increasing reliance on third-party libraries and services has escalated data privacy and security concerns among users. However, existing research does not systematically or standardize the assessment of these dependencies privacy implications.
Recognizing the critical need for data protection and the absence of a specific evaluation framework for third-party dependencies, this study aims to fill the gaps in the current literature by developing a robust PIA methodology. The methodology involved examining privacy policies and data access behaviours of third-party dependencies using Natural Language Processing (NLP) techniques to identify potential privacy risks.
The research made valuable contributions, such as creating a detailed framework for Privacy Impact Assessments (PIA) and establishing a risk rating score. This score offers valuable insights for safely integrating third-party dependencies. Ultimately, this research aims to enhance data protection and user privacy within the software development lifecycle. The findings of the completed research contributed significantly to enhancing privacy and security practices concerning third-party dependencies in software development
Enhancing Efficiency of Machine Learning Techniques with Feature Selection and Hyperparameter Tuning for Intrusion Detection towards Leveraging Cybersecurity
No full textCybersecurity enhancement is a continuous process due to increasing cyberattacks of late. Traditional security mechanisms were based on certain heuristics that could detect intrusions based on their detection process. However, with the emergence of Artificial Intelligence (AI) methods such as machine learning (ML) approaches, learning-based models are found efficient due to their ability to learn from labelled data continuously. It is found in the literature that ML models based on supervised learning show deteriorated intrusion detection performance when training samples are not with designed quantity and quality. Therefore, it is important to leverage performance of ML models with certain optimizations. This is the motivation behind this research which is aimed at building an intrusion detection system based on ML models with feature engineering and hyperparameter tuning optimizations. The system is evaluated using CICIDS2017 dataset. Intrusion detection system is implemented using binomial classification and also multi-class classification. In the binomial classification highest accuracy is achieved by RF model with 99.87%. In case of multi-class classification without optimizations, highest accuracy is exhibited by RF with 99.44%. In case of multi-class classification with optimizations, highest accuracy is exhibited by XGBoost with 99.64%
Using Genetic Algorithms for Optimized Feature Selection in Machine Learning and Deep Learning Models to Detect Phishing Websites
No full textGiven the advancement of network technology and the global rise in internet usage, cybersecurity concern has also risen. Phishing attacks which use misleading tactics to leak sensitive information bring about significant financial losses, reputational harm, system disruptions and legal consequences for both victim and organisation. This growing threat demonstrates the critical need for more robust detection systems, especially when accounting for evolving attack tactics. This study addresses the issue by investigating the use of a Genetic Algorithm (GA) combined with Support Vector Machine (SVM), Random Forest Classifier (RFC), Gradient Boost Classifier (GBC) and Graph Neural Network (GNN) to identify phishing websites. The findings of the experiment show that hybrid models perform better than single-base models, providing higher accuracy, F1-score, and AUC on a small dataset. According to this research, the GA-GNN and GA-RFC were the best and most improved model with accuracy of 95%, F1-score of 94.74% and AUC of 95.45%. The paper does admit several limitations such as, performance drop with larger sample size. Nonetheless, the study highlights the promise of hybrid models in tackling phishing detection problems and raising detection precision