جستجوی مقالات مرتبط با کلیدواژه "resampling" در نشریات گروه "برق"

The production of fake news were practiced even before the advent of the internet. However, with the development of the internet and traditional media giving way to social media, the growing and unstoppable process of making and spreading this kind of news have become a widespread concern. Fake news by disrupting the proper flow of information and deluding public opinion, potentially causes serious problems in society. Therefore, it is necessary to detect such news, which is associated with some challenges. These challenges may be related to various issues such as datasets, events, or audiences. Lack of sufficient information about news samples, or an imbalance are the main problems in some of these datasets, which will be addressed in this paper. In the proposed model, firstly the key features in relevant datasets will be extracted to increase information about news samples. After that, using the K-nearest neighbors, a genetic, and TomekLink algorithms as the cleaning techniques, as well as designing a Generative Adversarial network, as a technique for generating synthetic data, three novel methods in the area of hybrid resampling will be presented to balance these datasets. The presented methods cause a significant increase in the performance of the deep learning algorithms to detect fake news.

Due to today’s advancement in technology and businesses, fraud detection has become a critical component of financial transactions. Considering vast amounts of data in large datasets, it becomes more difficult to detect fraud transactions manually. In this research, we propose a combined method using both data mining and statistical tasks, utilizing feature selection, resampling and cost-sensitive learning for credit card fraud detection. In the first step, useful features are identified using genetic algorithm. Next, the optimal resampling strategy is determined based on the design of experiments (DOE) and response surface methodologies. Finally, the cost sensitive C4.5 algorithm is used as the base learner in the Adaboost algorithm. Using a real-time data set, results show that applying the proposed method significantly reduces the misclassification cost by at least 14% compared with Decision tree, Naïve bayes, Bayesian Network, Neural network and Artificial immune system.