Using NaïVe Bayes Classifier to Indentify Falsified Text Information

The present study investigates the problem of spreading fake text news within the framework of AI-driven generative technologies. This research builds upon prior works into methodologies for categorizing such data within contexts marked by societal instability. The naïve Bayes classifier was adopted as the primary method, leveraging a feature set derived from an analysis of deceptive news attributes. To evaluate model performance in terms of both speed and accuracy, a method employing linear additive convolution with weighted coefficients was employed. A hybrid neural network, developed in preceding research, served as a benchmark for comparison. The findings indicate that the proposed probabilistic approach exhibits superior effectiveness. Nonetheless, given the specific nature of this comparison, these outcomes lay groundwork for further exploration into the applicability of the chosen model family.