PHISHING URL DETECTION USING RESIDUAL PIPELINING

Abstract

Phishing websites are the ones with same name and appearance as an offi cial website. Nowadays, URL phishing is increasing and are causing great threats, they pretend to be the same as the official website and steal user information. Earlier blacklists were used to identify these URLs; however, they cannot identify the one-time Uniform Resource Locators (URL). Deep learning methods are employed to avoid such risks and also to improve the detection accuracy and decrease the misjudgment ratio. In URL detection using residual pipelining, the common URL features along with some senti mental values are extracted and fed onto a residual pipeline. The residual pipeline consists of convolutional blocks and inverted residual blocks. The result obtained from this block is finally passed onto the MLP output layer where the actual detection and classification of an URL takes place. The dataset is obtained from Kaggle. The accuracy, precision, F1 score, and re call are being measured and it is observed that it is much higher than those of several other traditional algorithms