The insights of this study can assist SME decision makers in adopting cloud ERP in their organisations. Further, the integration of TOE and UTAUT frameworks in a theoretical model represents an innovative approach, making a significant contribution to the existing body of knowledge.
Eve teasing, a form of public harassment and assault
on women, is a significant issue that causes severe distress,
particularly among young women and girls. The rising incidence
of eve teasing in Bangladesh has led to severe crimes such as
rape and murder, with many offenders escaping due to a lack of
evidence and effective monitoring. This paper presents a novel
approach using computer vision and machine learning methods
to detect eve teasing from video material in various situations.
Our proposed solution combines gender detection, expression
analysis, and gesture recognition to identify behaviors indicative
of eve teasing. The system integrates male-female identification,
human behavior detection, and CCTV-based monitoring or video
footage analysis to identify such critical incidents. Additionally,
the approach includes determining the participants involved in
the scenario to provide comprehensive evidence of harassment.
By enabling more accurate detection and verification of eve
teasing in real time, our method offers a promising tool to
help victims prove harassment and support law enforcement in
apprehending offenders, thereby contributing to a safer public
environment.
This study addresses significant data privacy issues in telemedicine by proposing a secure method for transmitting sensitive medical information over unsecured networks. The novel pipeline combines state-of-the-art image classification with image steganography to integrate patient diagnosis and personal information into medical images, ensuring data privacy and protection from unauthorized access. The approach uses fine-tuned convolutional models for classification and an encoder-decoder architecture for steganography, safeguarding the transmission of medical data while maintaining image quality.
1. Conducted multi-class classification tasks to identify distinct attack types using various machine learning algorithms.
2. Evaluated the use of Hardware Performance Counters (HPC) and kernel events as features, both individually and in combination, to compare and analyze the performance of these algorithms.
3. Extensive experiments using ten-fold cross-validation demonstrated that Random Forest based machine learning model achieved the highest overall accuracy of 93.4%.
4. Attack classes comprising more than 20% of the samples attain nearly 100% accuracy, while classes with less than 3% samples tend to underperform.
This paper explores the required capacity of renewable and storage resources for an isolated grid in long-term planning. The study also examines the impact of Dunkelflaute events on capacity planning and demonstrates how a balanced mix of variable renewable energy can mitigate network challenges.
A novel adaptive feature extraction technique is proposed in partitioning
to overcome the problems of over-tracking and under-tracking. Mean spectral kurtosis analysis is performed across several partitioning techniques to assess their symbolization effectiveness. This enables the selection of the most appropriate partitioning technique. This enhances the localization and tracking of target objects by focusing on robustness to noise and multipath effects. The proposed method learns and estimates the distance range simultaneously, thereby eliminating the need for a separate offline training phase and the storage of reference coordinates. Experimental results using LoRa highlight the proposed method’s efficacy in real-time localization, tracking, and superiority over the state-of-the-art method.
The goal of this project is to predict customer behavior from a large real-world e-commerce dataset using tree-based machine learning modeling techniques that will employ decision tree, random forest, and gradient boosting. Each of the models will be evaluated and compared to determine which of the three is the best model for predicting customer behavior.
The goal of this project is to predict customer behavior from a large real-world e-commerce dataset using tree-based machine learning modeling techniques that will employ decision tree, random forest, and gradient boosting. Each of the models will be evaluated and compared to determine which of the three is the best model for predicting customer behavior.