Chrome Extension For Malicious URLs detection in Social Media Applications Using Artificial Neural Networks And Long Short Term Memory Networks
Saumya Shivangi, Pratyush Debnath, K. V. Sajeevan, Annapurna Dammur · 2018
Social media applications have become an integral part of our life, business and society today. Due to their increasing number of users and growing popularity, many organisations use them as a medium to generate income. The businesses use social media analytics and advertisements to increase their revenue. Although social media applications were initially built to connect people across the globe, they have now turned into one of the most favoured ways of propagation of cyber crimes. Most users lack cyber awareness and fall prey to the malicious activities distributed via Uniform Resource Locators (URLs) and advertisements. When a user visits the malicious URL, it makes the hackers privy to a lot of personal and sensitive information of the user. To overcome the problem of malicious URLs victimising users we propose a tool deployed as a chrome extension. This tool, analyses URLs and classifies them using two different neural networks, Artificial Neural Networks (ANN) and Long Short-Term Memory (LSTM) networks which is a specific type of Recurrent Neural Network (RNN). The major objective of the proposed model is to aid the users to avoid becoming a victim of malicious and fraudulent activities like malicious URLs, phishing and social engineering that favour social media as their target medium by detecting them accurately. The model proposed is scalable, easy to train and compatible with devices of varied hardware specifications. The proposed model gives excellent accuracy and overcomes several issues faced by the existing systems.