Optimized Convolution Neural Network Based Fake News Detection Using Sentiment Analysis

In the era of social media, the social media and smartphone popularity has enhanced exponentially. By the electronic media, fake news has rising quick with new information which are hugely untrustworthy. The search engine like google are incapable for fraudulent of news because its limitation with restricted keywords. The Optimized Convolution Neural Network (OPCNN) is proposed for classifying fake news into actual and fraudulent news based on sentiment analysis. which accommodate varying complexity levels by adjusting its architecture in Fake News Detection (FND). The Principal Component Analysis (PCA) is used to extract features from preprocessed images which reduces the data dimension comprising numerous related variables and recalls the high change in real data. The ISOT dataset is preprocessed through four various techniques. The recall, accuracy, f1 score and precision with ISOT dataset are considered to evaluate OPCNN performance. The OPCNN realizes 99.58% recall, 99.67% accuracy, 99.61% f1score and 99.64% precision for ISOT dataset when matched to previous techniques like Random Forest (RF) and Sea Turtle Foraging Optimization-based Fake News Detection and Classification (STODL-FNDC).