8333 research outputs found
Sort by
Enhancing CAN Bus Security Using Message Authentication
No full textThe Controller Area Network (CAN) protocol, introduced by Robert BOSCH GmbH in 1983, is a critical component of automotive communication and other domains. This thesis aims to improve the security of the CAN bus by implementing message authentication mechanisms. With modern vehicles containing around 100 Electronic Control Units (ECUs) responsible for critical functions, a reliable communication network is essential. The CAN bus includes advantages such as high immunity to electrical interference and easy wiring, which makes it perfect for the automotive industry.
This research addresses the security challenges within the current CAN bus architecture, with a specific focus on message authentication, and explores how its implementation can improve the overall security. The study includes an extensive literature review of the CAN bus, its security challenges, and existing research on message authentication. Additionally, the methodology section details the setup of the research environment, including the tools and techniques used to investigate the research question.
The findings demonstrate that while implementing Hash-Based Message Authentication (HMAC) can significantly improve the security of the CAN bus communication in vehicles it also has a negative impact on the latency of the network. The study also highlights the need for further research into the security of the CAN bus, especially with the increasing number of ECUs and the growing communication of modern vehicles with back-end servers and other cars
Impact of the high-frequency public transport on the performance of the Machine Learning model for predicting the rental price in Dublin
No full textIn years the rising population, in cities has made it increasingly important to find affordable rental houses that match their true value. To address this concern our research employs a machine learning technology to predict rental house prices in the Dublin housing market. By acknowledging that multiple factors influence housing prices we adopt an approach that combines algorithms to enhance model accuracy and prediction. This study incorporates machine learning algorithms such, as Decision Tree, Gradient Boosting and K-Nearest Neighbour(KNN). This study predicts the prices of residential rental properties using the data from RTB (Residential Tenancies Board) which is open source data. The dataset has various features like location, property features etc. Additional features have also been added to the dataset by looking into the connectivity of the rental property with the high-frequency public transportation of Dublin. The high-frequency public transportation in Dublin is the luas and the darts. The luas have two different lines one is the green line luas, and another is the red line luas. We will also add more features like a number of luas and dart stops and also identified and then classify the postcodes of Dublin into North and South of Dublin. This study aims to provide accurate rental price prediction and also explain the connection between the various factors. After applying the regression machine learning algorithm, we also predicted the rental house price from the year 2023 to 2026 by applying the ARIMA Model Grid Search and Forecasting
Synch-Ur-Life: Technical Report
No full textThis project aims to create a straightforward, efficient task management system designed to improve productivity and organizational skills. It focuses on delivering a simple and effective way for users to manage their tasks, emphasizing ease of use, clear interfaces, and reliable functionality. The goal is to offer a solution that helps users streamline their daily routines, enhancing their ability to organize, prioritize, and complete tasks effectively. By providing a tool that is accessible and adaptable, the project seeks to support users in their personal and professional task management, making it easier for them to stay organized and focused
Impact of HRM on employee retention
No full textThe research has aimed to analyse the impact of human resource management on employee retention in the banking sector of India. The study has adopted a primary quantitative method through a survey. The data analysis has been done using MS Excel tool. The data analysis chapter ensures better collection of the data set regarding critical analysis over HRM Impact in the Indian banking Sector. Due to this, quantitative data is being designed accurately to get authentic information and an in-depth interpretation of the research topic. The study may provide valuable insight to HR professionals, executives, and legislators in the organisation seeking to improve employee retention tactics in the Indian banking industry by examining the challenges and barriers evaluating the successful application of HRM. The Indian banking sector is changing quickly due to several opportunities in client preferences, Technology, and regulations on the improvement of employee retention approaches. Additionally, the research paper's conclusion could provide proof of an HRM strategy tailored to the particular needs and features of the banking industry's improvement, enabling the stability and long-term growth of the organisation while preserving staff retention. The findings have made sure that employee recognition also significantly contributes to increased job satisfaction and morale. The chapter 5 is basically about the discussion part which phases of the analysis based upon the reviews of literature and the finding section based upon the impact of the HRM role of the employee retention in the workplace. The section basically focuses combining objectives which the research gathers from the finding section and improvises in it for the better understanding in the current research. Chapter 6 is essential for summing up the overall study as it states that the research has crucially developed multiple theories in the region of HRM and the retention of employees by amalgamating the “Expectancy Theory and Social Exchange Theory” into the evaluation. Moreover, by utilising these theories, the study has conducted a stronger acknowledgement of the motivational factors that regulate the retention of employees and highlights the feasibility of reinforcing the supportive environments of the work that recognize the contribution of the employees. In this way, this analysis has helped in expanding the insights into the sociological as well as psychological theories, which need to be incorporated into the practices of HRM, leveraging a more meticulous acknowledgement of the employee retention aspects within the organizations
Comparative performance of RF and GBM for short-term customer segmentation forecasting
No full textThe research explores the use of Recency, Frequency, and Monetary (RFM) analysis for customer segmentation. Companies often use segmentation techniques to generate insights on purchase behavior to quantitatively rank and group customers for targeted marketing campaigns. A typical question analysts face is how much data they need to perform those analyses and the confidence of predictions. However, little is known about what month is the easiest one to classify. Or, what is the best month in which the largest share of target customers is found? Here we present a detailed analysis of 1-month customer segmentation forecasting using two well-known machine learning techniques. We critically evaluate the classification accuracy on three segments, for “good”, “medium” and “bad” customers.
Based on the findings, we also evaluate the need for automated model selection and hyper-parameter optimization of the customer segmentation models. While many papers would go straight into the optimisation part, we want to verify whether such a range of tools is actually needed for improved classification accuracy. Based on the literature review, Random Forest and GBM classifiers were singled out as the top techniques for this classification task. It is however not known which one will deliver the best classification accuracy. As part of the research, we will show that plain RF and GBM are unsuitable for the task, as the distribution of “good” customers is uneven. To avoid this class imbalance, we used stratified classification.
The findings of the study point toward that, although April is the best month for prediction with RF, discussions have highlighted issues such as the classification of a few “good” customers and small sample sizes, and the need to tune the RFM score function or the “good” customers threshold. Besides, GBM outperforms RF by far, especially in those months that have a smaller number of “good” customers. GBM and RF give 98.5% and 91% classification accuracy, respectively with stratified classification using 10-fold cross-validation. Models are good enough, well over 90% accurate, and hence there is no need for further boosting of hyper-parameters. The results and research methodology are expected to provide valuable insights for analysts planning to do customer segmentation and forecasting of customer behaviou
An MLHOps-Driven Vision Transformer Approach for Pneumonia Classification in Chest X-Rays
No full textDiagnostic innovations are needed to enhance early detection and treatment of pneumonia, a global public health crisis. This work improves pneumonia detection from chest X-rays using Machine Learning Operations (MLOps) and Vision Transformers (ViT), a sophisticated deep learning (DL) model used in many computer vision applications. Pre-processed chest X-rays are given to a ViT model for feature extraction. The ViT encoder extracts hierarchical visual cues, whereas the classifier predicts pneumonia. The proposed method is tested using 5863 pneumonia-labeled NIH Chest X-ray pictures. The experiment compares the ViT model to a CNN classifier model for pneumonia classification based on accuracy, sensitivity, specificity, and AUC score criteria and it was found that ViT performed in terms of validation accuracy of 95.22% better than CNN accuracy of 94.84%. The ViT model execution was completed much faster than CNN execution (3018 seconds vs 7816 seconds). Based on these results, ViT was chosen to be implemented using MLOps practices for model training, evaluation, and deployment on Microsoft AzureML cloud. The suggested pneumonia detection on ML Health Operations (MLHOps) infrastructure using integrated ML pipelines allows rapid iteration and model optimization and ensures reproducibility for additional medical image analysis applications
Enhancing Cybersecurity through Comprehensive User Training Programs: A Study on Mitigating Social Engineering Threats
No full textIn order to successfully mitigate the growing threat posed by social engineering attacks, this study sheds light on the importance of user training programs specifically focused on the behavioural changes, confidence levels, threat perception and decision-making patterns.
A mixed-methods approach was adopted, engaging 30 participants from various cybersecurity backgrounds. Utilizing purposive sampling, a wide range of experiences was captured to understand the effectiveness of cybersecurity training. Both quantitative surveys and qualitative simulated social engineering attacks were combined to assess participant’s vulnerabilities and the impact of the training program, before and after its implementation
Android Malware Detection using Machine Learning and Convolutional Neural Network
No full textThe fast evolution of Android mobile devices has opened new possibilities for convenience and usefulness, but it has also given rise to Android malware, which poses a major danger to user security such as data theft. To discriminate between benign and malicious apps, effective detection systems are essential. Machine learning and Deep learning have emerged as a potent connect in this domain, capable of swiftly assessing Android application packages (APK files) and properly classifying applications. In this study, we examine the efficacy of various machine learning models and a deep learning model, with the Random Forest, Extra Tree, XGBoost, Stacking Classifier and Convolutional Neural Network (CNN), in detecting Android malware also Scikit Framework has been used for detecting malware in this research. Between these models, CNN stands out as the best performer, with respect to its precision, memory, and F1 scores to determine their correctness and dependability. This superior result demonstrates its capacity to reliably detect and categorise Android apps, emphasizing its importance in improving mobile security. This report offers critical insights into the state of Android malware detection using machine learning, offering a path forward to enhance Android security and shield users from potential security threats in the ever-evolving landscape of mobile technology
Enhancing Malware Detection Using Stacked BiLSTM with Attention Mechanism: A Deep Learning Approach
No full textThe detection of malware offers quite a few challenges due to the ever-evolving nature of threats in the field of cybersecurity obviously. Traditional methods of malware detection methods already struggling to keep pace due to the rapidly changing landscape of malicious software. This study introduces a novel approach with the help of deep learning techniques which will enhance malware detection efficiency and address the prior work limitations like traditional methods which are already facing challenges. There are four models which are going to be performed in this project i.e., Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Memory (BiLSTM) networks and Stacked BiLSTM with an Attention Mechanism. These models are more effective in identifying complex patterns and dependencies found in Portable Executable (PE) files, which are frequently used as vectors for the spread of malware. Using an extensive dataset that includes both malware and benign PE files that are sourced from reliable repositories, this study guarantees a strong training framework. While this project tests all the four models, and after comparing the results, stacked BiLSTM with attention mechanism wins the accuracy margin. This study fills a critical gap in malware detection by leveraging advanced machine learning techniques to enhance cybersecurity defenses and mitigate the risks posed by evolving malware threats
Enhancing Stock Market Forecasting on the Bombay Stock Exchange through an Evolutionary Approach and Deep Learning Techniques
No full textIn addressing the research problem of improving BSE stock future prediction through deep learning models. the deployment of techniques for analyzing and forecasting stock prices on the Bombay Stock Exchange utilizing deep learning signifies a major leap forward from conventional approaches. By adeptly handling vast as well as complex dataset and uncovering intricate patterns within them, these advanced AI methodologies offer a more nuanced and accurate prediction of market trends. This not only improves investment decision-making but also promises to elevate the overall efficiency and stability of the market. As such, embracing deep learning in financial analysis on the BSE is pivotal for investors and analysts seeking to navigate the complexities of the modern financial landscape effectively. In the focused study on the prediction of Bombay Stock Exchange (BSE), LSTM, and GRU strategically employed to analyze market data, By tapping into their skill in capturing complex sequential patterns that are essential for predicting time series, we've harnessed the power of LSTM's ability to learn long-term dependencies and GRU's proficiency in handling intricacies inherent in time series data. The result has been remarkably successful in accurately forecasting the near-term movements of the stock market. The application of these complex models resulted in a significant improvement of predictive accuracy, showcasing their superiority over conventional methods. This breakthrough not only facilitates more informed investment decisions and effective risk management but also highlights the potential of advanced deep learning techniques in navigating the dynamic nature of financial markets, marking a notable stride in the realm of financial forecasting and the development of trading strategies