Design of SMOTE Analysis-Based Phishing Attack Detection Scheme for Web of Things Environment

As the world becomes more digital, cyber security is crucial for individuals, groups, and businesses. Secure communication must also be provided in the Web of Things (WoT) environment. Protecting its data against fraud and loss is essential. Phishing attacks, which deceive users to steal sensitive data, continue to threaten cyber security. This paper introduces SPADS-WoT, a machine learning-based phishing detection method using Synthetic Minority Over-sampling Technique (SMOTE) analysis. The SPADS-WoT uses a UCI Machine Learning Library dataset with a balanced distribution of authentic and phishing URLs from the SMOTE study. Logistic regression, random forest, SVC, K-means, and naive Bayes are used to train and evaluate the model. Our experiments show that SPADS-WoT is successful, with Random Forest getting the highest accuracy score of 99.95%$$ 99.95\% $$. We also compare SPADS-WoT to other systems. SPADS-WoT surpasses other methods. The SPADS-WoT might counteract phishing attempts by using machine learning automation to improve cyber security detection accuracy and efficiency.