A Novel Framework for Fake News Detection Using LDA and QDA

In today’s digital age, the rise of online media has made information readily accessible, but it has also facilitated the rapid spread of fake news. This phenomenon undermines trust in reliable news sources and fosters confusion among the public. Consequently, individuals must exercise caution when sharing information, verifying its credibility to prevent the spread of false information and uphold the trustworthiness of reputable channels. A highly accurate system has been designed to detect fake news efficiently and with a high level of accuracy. In this paper, an enhanced approach is suggested for classification based on the Linear Discriminant Analysis and Quadratic Discriminant Analysis models, some advanced methods that help fight against fake media stories. A number of natural language processing techniques, including tokenization, stop word removal, and URL and HTML tag removal, were applied to a comprehensive dataset of fake news collections. For the purpose of classification, two models-Linear Discriminant Analysis and Quadratic Discriminant Analysis models-were applied for the correct detection of fake news. In order to evaluate effectiveness, performance measures such as accuracy, precision, recall, F1 score, confusion matrix, and ROC curve were critically analyzed. The accuracy of LDA is 97%, while QDA achieves 67%. This paper attempts to provide a step toward efforts engaged in fighting misinformation in the digital age and trying to keep reliable news sources intact.