Online Social Networks (OSNs), fake identities, machine learning algorithms, random forest algorithm, Naïve bias.
Abstract
Social media is one of the preferred communication platforms and has become a target for spammers and scammers. Instagram is a prominent Online Social Network platform for users. It has a lot of features for posting pictures, videos, and text. It created a wide range of communication all over the world. But it also harms people. It creates a way for attackers to steal personal data from individuals, spread rumors and fake news, defame someone’s character, cyberbully, and misdirect people to fake websites. To prevent this type of attack there are a few methods that follow naive Bayes to detect those fake identities. The model can detect fake identities by using attributes like profile pictures, number of followers, number of followers, and the content that they use for chats. The existing methods consist of low accuracy rates, high complexity, and require skilled persons. Random forest has features like the ability to perform both regression and classification, searching for the best, and logistic algorithms have features like containing low bias, higher variance, and more collinearity. By combining these powerful algorithms, we can increase the accuracy rate, and decrease the complexity and there is no need for skilled persons to detect those fake identities. By this process, we can easily find an account that is genuine or fake with an accuracy rate of 91% which is more accurate than the previously existing methods that follow Navie bias algorithms.
Article Details
Unique Paper ID: 161533
Publication Volume & Issue: Volume 10, Issue 4
Page(s): 495 - 499
Article Preview & Download
Share This Article
Join our RMS
Conference Alert
NCSEM 2024
National Conference on Sustainable Engineering and Management - 2024