Modern social media dominates many areas. Social media use has skyrocketed. Social media helps us connect with people and express ourselves. This enabled identity fraud, fact falsification, and other assaults. Recent evidence reveals that social media accounts outnumber active users. This suggests a recent spike in fake accounts. Social media networks struggle to identify fake accounts. Due to the rise of fake accounts, advertisements, etc. on social media, detecting them is crucial. Traditional methods fail to distinguish real from fake accounts. The aforementioned articles are outdated due to advances in fake profile creation. The new models employed automated posting and commenting, false facts, and promotional material in spam to identify fake accounts. Multiple algorithms, each with their unique traits, are needed to combat fake accounts. Naive bayes, support vector machine, and random forest no longer identify fake accounts. This work provided an innovative technique to address this problem. Three-attribute decision trees and gradient boosting were used. Engagement, phony activity, and spam comments are examples. Machine Learning and Data Science helped us uncover bogus profiles.