Challenges in implementing machine learning (ML) include expanding data resources within the finance sector. Banking data with significant financial implications are highly confidential. Diverse breaches and privacy violations can result from a combination of user information from different institutions for banking purposes. To address these issues, federated learning (FL) using a flower framework is utilized to protect the privacy of individual organizations while still collaborating through separate models to create a unified global model. However, joint training on datasets with diverse distributions can lead to suboptimal learning and additional privacy concerns. To mitigate this, solutions using federated averaging (FedAvg), federated proximal (FedProx), and federated optimization methods have been proposed. These methods work with data locality during training at local clients without exposing data, while maintaining global convergence to enhance the privacy of local models within the framework. In this analysis, the UNSW-NB15 and credit datasets were employed, utilizing precision, recall, accuracy, F1-score, ROC, and AUC as performance indicators to demonstrate the effectiveness of the proposed strategy using FedAvg, FedProx, and FedOpt. The proposed algorithms were subjected to an empirical study, which revealed significant performance benefits when using the flower framework. Consequently experiments were conducted over 50 rounds using the UNSW-NB15 dataset, which achieved accuracies of 99.87% for both FedAvg and FedProx and 99.94% for FedOpt. Similarly, with the credit dataset under the same conditions, FedAvg and FedProx achieved accuracies of 99.95% and 99.94%, respectively. These results indicate that the proposed framework is highly effective and can be applied in real-world applications across various domains for secure and privacy-preserving collaborative machine learning.

The 21st century has seen a significant increase in fake news, mainly due to the rise of social media and greater internet access. The rapid online spread of fake news poses a significant challenge, as traditional methods of fake news detection are struggling to catch up. To address this, our study centered on developing reliable and empirically validated data-driven algorithms for fake news detection. Through a detailed evaluation of both traditional and advanced machine learning models on datasets, our results show that while traditional models exhibit better generalization across varied datasets, deep learning models, particularly transformer-based architectures, achieve higher accuracy in detecting nuanced textual patterns. The goal of our research is to shed light on the strengths and limitations of distinct model architectures, providing insights to their capacity to generalize across various datasets and adapt to different task needs. Though traditional models frequently exhibit better generalization capabilities, the ideal model choice can depend on particular task and dataset distribution. Systematic fake news detection has the potential to preemptively halt the spread of false information, thus alleviating its damage on society. By utilizing evaluations from both traditional and deep learning models, our study aims to develop an extensive approach to fake news identification. The fundamental aim is to develop a model that shows high accuracy, robustness and generalizability across different datasets and information frameworks. Through this study, we seek to improve systematic fake news detection systems, contributing to ensure the integrity of information flow and democratic functions in the digital age.