8333 research outputs found
Sort by
What, When, and How of Responsible Leadership: Taking Stock of Eighteen Years of Research and a Future Agenda
No full textBecause research on responsible leadership has grown significantly in recent years, we conducted a systematic review of research on responsible leadership. Our overall goal was to establish a comprehensive understanding of alternative definitions of responsible leadership, its theoretical foundations, and distinctions from other moral leadership constructs. Drawing from 194 studies, we first clarify the conceptual underpinnings of responsible leadership, and how it differs from other constructs in the moral leadership domain, thus highlighting its value as a construct. Second, we identify and evaluate the prominent theoretical frameworks that underpin responsible leadership. Third, we conceptualize the antecedents, mediating factors, contingency variables and outcomes of responsible leadership. Fourth, we offer important recommendations for future research that will move the field forward. Overall, our review provides insights to advance an understanding of responsible leadership
Opinion mining on newspaper headlines regarding the US elections using NLP, SVM and Deep Learning
No full textThis research investigates sentiment analysis on newspaper headlines concerning the 2024 U.S. Presidential Election using Natural Language Processing (NLP),Support Vector Machine(SVM). Multiple researches have been done in opinion mining for online blogs, Twitter, Facebook etc. using the public social media platforms but in this paper we are focused towards the headlines which first attracts the consumer to further read the content. The primary objective of the research is to predict the public sentiment and its potential influence on electoral outcomes by analyzing the important headlines from major news outlets. The study initially utilized Support Vector Machines (SVM) with TF-IDF vectorization with the further refinement was undertaken by incorporating Word2Vec embeddings with an improved accuracy.
To enhance performance and to understand the small nuances in the findings advanced transformers like BERT and RoBERTa were explored, leveraging their pretrained architectures for fine-grained sentiment classification. Despite the moderate gains with using the transformers, the results highlighted the inherent challenges of sentiment classification in nuanced, politically charged content. The project focused on the early stages such as feature engineering and preprocessing techniques, such as Named Entity Recognition (NER), to contextualize sentiment further
House Price Prediction in Beijing
No full textReal estate price in particular, and house price specifically, has been an important area of research as it aims at predicting forecast that assists the side stakeholders in the decision-making process of the property. Such old models as Linear Regression, Decision Trees, and Random Forests were applied to predict house prices a long time ago; however, they do not handle well second-order effects and nonlinear interactions of features that are always present in real-life data. Even these models have also come with the problem of causing overfitting and do not generalize properly to new data sets. To overcome these challenges, this study will incorporate ordinary learners and complex algorithms of XGBoost and ANN learners. The strategy of this work involves comparing six models which include Linear Regression, Decision Tree Regressor, Random Forest Regressor, Gradient Boosting, XGBoost Regressor, and Artificial Neural Network. Among these models, the performance of XGBoost was the highest since it attained an MSE of 2,873.27 and R² of 0.947 indicating that this model can easily identify intricate patterns as well as interaction. The ANN also demonstrated good results achieving MAE = 33.53 as well as R² = 0.9344 while stressing its flexibility. As this research shows, with enhanced artificial learning methods, the prediction of house prices can be made more accurate and reliable than with the regular approaches
Enhancing Cryptocurrency Price Prediction using Transformer-Based Models for effective Time-Series Analysis
No full textCryptocurrency prices highly fluctuating and accurate price prediction of cryptocurrency is difficult but important task for the investor or trader. The conventional machine learning models are not well equipped to capture the sequential and temporal relationships which are present in cryptocurrency time-series data. This work aims at analyzing the of using transformer-based models that are good at modeling long-term dependencies and intricate structures for predicting cryptocurrency prices. The emphasis is on the method which helps to forecast the Bitcoin prices based on the historical prices and other characteristics of the market. The forecasting performance of the transformer model is compared with other baseline models such as SVM, Gradient Boosting (GBM), Random Forest (RF), RNN and LSTM. Performance metrics such as MSE, MAE, RMSE, and R² are applied to compare the accurate predictions of the model. The transformer model outperformed all other models with an MSE of 123709.59, MAE of 295.12, RMSE of 351.72 and R² of 0.9801. In comparison, traditional models such as SVM, Random Forest, Gradient Boosting and deep learning models like RNN and LSTM are unable to recognize the long-term dependencies and patterns in the change in the price of cryptocurrencies. These results demonstrate that the transformer model outperforms other models in the forecast of highly unpredictable cryptocurrency prices and positions them as a potentially viable solution for enhancing the accuracy of financial predictions. It also highlights directions for future research, such as the including of the other features in the market, real-time prediction models, and more comprehensible models, to aid both theoretical and applied research in cryptocurrency price prediction
Enhancing IoT Security through Anomaly-based Intrusion Detection Systems
No full textThe advancement of the Internet of Things (IoT) has seen rapid growth in the industrial connectivity and automation rates considerably. However, this growth has also created important concrete cybersecurity threats as IoT networks are now in the crosshairs of highly developed cyber attacks. The first problem is that, unlike more traditional networks, the emerging IoT networks exhibit high levels of heterogeneity and low available resources; and the second is that most current IDSs have rigid architecture and are not suitable for the IoT networks. This work offers an anomaly-based IDS for improving the security of IoT networks that exploits state-of-the-art ML and DL methodologies. The proposed system includes Gradient Boosting Machine (GBM), k-Nearest Neighbour (KNN), and Naive Bayes with Graph Neural Networks (GNNs): Graph Convolutional Networks (GCNs) and Graph Isomorphism Networks (GINs). In results of experiments, Random Forest and KNN surpass competitors with such diagrams as 96.30% and 98.19% correspondingly, while GNNs are also combined with GIN and give excellent results in resect of complex traffic pattern detection with 79.12% of accurate classification. These results prove that hybrid anomaly-based IDSs are useful to achieve a steady and efficient IoT cybersecurity model
Predictive Modelling Coronary Artery Disease and Hypertension Using Machine Learning
No full textA Diseases related to the heart such as Hypertension (HT), and coronary artery disease (CAD) are major global health hazards. The timely prediction can help in performing preventive measures, thus providing a better patient outcome. Present models. Current techniques, despite their progress in predictive modeling, are often inapplicable for generalizing across complex and heterogeneous patient populations. This limitation has reduced their accuracy and reliability in real-world clinical settings, indicating that there is a need for more robust models that could address these issues and provide better prediction performance. In this study, by overcoming the data imbalance issue and utilizing ensemble methods with the data balancing through CTGAN, developed machine learning models that can predict the HT and CAD correctly with high accuracy. HT model generalized well over datasets attaining a test accuracy of 97% with balanced precision and recall. Test accuracy for the CAD model: 92% with a recall of 0.91 of CAD-positive cases, meaning it is able to reliably classify patients at risk of CAD. The CAD model shows minimal overfitting, with the training accuracy at 94%. The findings suggest that balancing the data can improve the accuracy to levels that can be clinically useful, and the ensemble model provides a reliable tool for accurate risk assessment for healthcare providers in the early stages of patient care. Larger datasets and advanced efforts over the model development on how to be more sturdy can be seen in future works
Potential Improvement in Sales of EVs through Effective use of Sales Trend Analysis in Strategic Decision Making
No full textThe automobile industry is gradually transitioning to EVs because the consumers have realized the importance of the environment and there are inventions in the industries that are employed in the vehicles. As for the Such an approach indicates that the market penetration of EVs during this period appears to have been augmented. to only 35% by the investigated automobile firm by the year 2023. The research goal of the present work is as follows: the analysis of the factors that may affect consumption, as well as the identification of measures to improve them. promotion. With this approach in defining sales trends, there is the need for research analysis. Interact with several products and services, customer’s behaviour, and conditions in the marketplace. to generate recommendations for EV vendors. The purpose of this research study is to be contributing to the stream of knowledge in the existing body of literature when it comes to gaining a great and exhaustive understanding of the actual sales pattern and consumers’ attitudes towards the promotion of the utilization of Regarding EV and providing concrete recommendations on how to boost the sales of the EV
From Traditional to Advanced Machine Learning: A Comparative Study of Political Tweet Sentiment Analysis
No full textThe growing influence of social media networks on political discourse requires advanced sentiment analysis to recognize better public viewpoints revealed in complicated and diverse textual data. Existing approaches typically struggle to stabilize computational effectiveness with the ability to capture contextual nuances in sentiment classification jobs. In this research, we investigate machine learning and deep learning strategies, for analyzing the sentiments using tweets. To evaluate high-dimensional data with complex linguistic patterns, we preprocessed the data and fine-tuned it for better results. The Outcomes suggest that while SVM attained an accuracy of 85.91% because of its performance in structured data LSTM outmatched a little with an accuracy of 86.34%, succeeding at capturing nuanced linguistic features. From these results, we can understand that LSTM is better suited for sentiment analysis
Enhancing Risk Assessment in Legal Documents through Advanced Machine Learning
No full textIn this thesis, we will conduct research to examine the use of BERT, Legal-BERT and Isolation Forest models to strengthen the risk assessment and anomaly detection of these legal documents. It shows that Legal-BERT, a BERT model pretrained on legal texts can outperform a general BERT model well beyond chance level with a test accuracy of 76.47%, and excellent specificity in precision, recall and F1 score measures. In addition, the Isolation Forest Algorithm, an unsupervised learning model, was able to find 975 anomalies among 19,501 clauses, and we can say that it is able to find deviants from normal threads of law. Most notably, real practice was the proving ground for these models, as they were tested in real-time on new legal texts and their applicability was confirmed. These findings support the opportunity advanced machine learning models provide for automation, increased accuracy, and scale in the analysis of legal documents and constitute a meaningful step toward automated risk assessments applied to real-world legal scenarios within the field of legal informatics
Comparative Analysis of Machine Learning Models for Mental Health Assessment Using Music Therapy
No full textThis project focuses on how machine learning can be used to estimate the levels of depression relying on the socio-economic and demographic variables, given the global rise of mental health disorders, such as, depression, anxiety, and OCD. The focus is in using individual data aimed to detect these issues at an earlier stage and deliver more tailored approaches that may help address these difficulties. KNN, Neural Networks, Decision Trees, Random Forests, and Gradient Boosting were identified as the models in the study. All these models were trained and assessed on a constrained structured dataset with prominent evaluation metrics including MSE and R² Score. The assessment exposed various findings regarding the prognostic capability and versatility of each model. Just like KNN, MLP has shown higher predictive accuracy with lowest MSE that makes this algorithm capable of capturing local data patterning. Neural Networks demonstrated the feature of capturing nonlinear relationship structure of the data. Some of the findings gives insights comparing many classifiers such as KNN outperforms the rest in accuracy and simplicity while Neural Networks outperforms in complex data features. On model selection, the project reminds the user of some dataset features and level of interpretability required when selecting suitable models. In doing so, this work creates a basis for further research to continue to build and refine the modeling of mental health issues and to extend the use of such modeling for practical endeavours in the early identification of possible disorders and the pursuit of appropriate intervention plans