A DEEP LEARNING BASED IMPROVED DETECTION FOR PHISHING ATTACKS IN SOCIAL ENGINES

The social engines used for Phishing attacks have become much more sophisticated, hosting the malicious components in a way that is nearly impossible for traditional detection mechanisms to pinpoint the threat accurately. An example of one of those tools in AI is deep learning, which has become so common in the world of AI since we can use it to nd out/counter such attacks. Select detection for phishing attacks in social engines. Automatic anti-virus monitoring software was created using deep learning, where neural networks were employed to interpret and learn the patterns of email content, social media postings, and webpage construction. The models are designed to detect features such as malicious URLs, obfuscated text, and sender identication that are typically indicative of phishing. Data collection is the rst step towards a well-functional working of this system. With phishing, that meant collecting actual emails and social network messages and webpages to train their neural networks. Pre-processing: Data is processed, we clean our dataset, and feature extraction is done in the next step. Now, as the next step, we have DL-based models like RNNs or CNNs trained with this preprocessed data to learn patterns that ag if it is a potential phishing attack.