Nowadays, the consequence of quick access to information has lead to the spread of fake news, which has a strong damaging impact on democracy, justice, and public trust. Thus, it is crucial to analyze and evaluate detection methods for fake news. This paper focuses on the detection of Romanian fake news. In this study, we made a comparative analysis of machine learning algorithms and Transformer-based models on Romanian fake news detection using three datasets—FakeRom, NEW, and both FakeRom + NEW. The NEW dataset was build using a scrapping algorithm applied on the Veridica platform. Our approach uses the following machine learning models for detection: Naive Bayes (NB), Logistic Regression (LR), and Support Vector Machine (SVM). We also used two Transformer-based models—BERT-based-multilingual-cased and RoBERTa-large. The performance of the models was evaluated using various metrics: accuracy, precision, recall, and F1 score. The results revealed that the BERT model trained on the NEW dataset consistently achieved the highest performance metrics across all test sets, with 96.5%. Also, Support Vector Machine trained on NEW was another top performer, reaching a very good accuracy of 94.6% on the combined test set.

Nowadays, the consequence of quick access to information has lead to the spread of fake news, which has a strong damaging impact on democracy, justice, and public trust. Thus, it is crucial to analyze and evaluate detection methods of fake news. This paper focuses on the detection of Romanian fake news. In this study, we made a comparative analysis of machine learning algorithms and transformer-based models on Romanian fake news detection using three datasets – FakeRom, New, and both FakeRom + New. The NEW dataset was build using a scrapping algorithm applied on the Veridica platform. Our approach uses for detection machine learning models: Native Bayes (NB,) Logistic regression (LR) and Support Vector Machine (SVM). We also used two transformer based models – BERT based multilingual based and RoBERTaLarge. The performance of the models was evaluated using various metrics: accuracy, precision, recall and F1-score. The results revealed that BERT model trained on NEW dataset consistently achieved the highest performance metrics across all test sets. Also, Support Vector Machine trained on NEW was another top performer, reaching a very good accuracy on the combined test set.

Classical machine learning (ML) models, including Logistic Regression (LR), Support Vector Machines (SVM), and Random Forests (RF), remain widely used in practical applications due to their efficiency, interpretability, and low computational costs. However, their security properties against different adversarial threats are often evaluated independently rather than within a unified comparative framework. This paper presents a unified empirical evaluation of LR, Linear SVM, and RF models across image (MNIST and CIFAR-10) and text (AG News) classification domains. The models are analyzed under three adversarial scenarios: training-time poisoning, inference-time evasion, and black-box model extraction attacks. The empirical results reveal architecture-dependent security trade-offs: while RF showed higher resistance against random label noise but remained vulnerable to targeted poisoning strategies. In contrast, linear models (LR and Linear SVM) exhibit higher susceptibility to black-box extraction, achieving replication fidelities up to 96.22% under active querying strategies, while showing more predictable performance degradation under the evaluated evasion settings. Furthermore, the study identifies an optimization-related effect in which gradient limitations may contribute to Linear SVM exhibiting higher apparent resistance against iterative attacks compared with single-step perturbations. The evaluation shows that model robustness depends on both the attack type and the underlying architecture. Random Forest achieved lower degradation in some poisoning scenarios, while linear models showed different sensitivity patterns under targeted attacks.

Classical machine learning (ML) models, including Logistic Regression (LR), Support Vector Machines (SVM), Random Forests (RF), and XGBoost, remain widely used in practical applications because of their efficiency, interpretability, and relatively low computational costs. However, their security properties against different adversarial threats are often evaluated independently rather than within a unified comparative framework. This paper presents a unified empirical evaluation of Logistic Regression, Linear SVM, Random Forest, and XGBoost models across image (MNIST and CIFAR-10), text (AG News), and tabular (Adult, Spambase) classification domains. The models are evaluated under three adversarial scenarios: training-time poisoning, inference-time evasion, and black-box model extraction attacks. The empirical results reveal architecture- and attack-dependent security trade-offs. Random Forest demonstrated relatively greater resilience to random label noise in several configurations but remained vulnerable to targeted poisoning, particularly on AG News. In contrast, Logistic Regression and Linear SVM achieved high black-box extraction fidelity under some active-query configurations, reaching 96.08% and 94.02%, respectively, on MNIST with Q = 10,000. Their evasion results were comparatively interpretable under the evaluated attack procedures, although the observed performance depended on the perturbation budget and optimization configuration. In particular, the higher apparent accuracy of Linear SVM under PGD than under FGSM should be interpreted as a configuration-dependent observation rather than evidence of intrinsic robustness or confirmed gradient masking. Overall, the findings indicate that model robustness depends on the interaction between the attack strategy, data representation, and model architecture.