Attention-based Deep Learning Model for Accurate Detection of Misleading News

With the inception of big data, there is a lot of information that goes unfiltered, contributing to fake news. This false information can lead to catastrophic results in the long run. Hence, to curb this spread, we aim to provide constructive techniques that can help detect this kind of misleading news. In this paper, we have introduced a novelty model, the GRU-Attention Network, and shown the comparison of the novelty model with existing machine learning algorithms like Naive Bayes, SVM, BERT, and DeBERTa. The analysis is based on the dataset, which was extracted from diverse sources and contains over 20,000 news samples. The proposed work results in an accuracy ranging from 85.7% to 99.9%, which is very good for classifying most of the information