Spam Detection with Complex-Valued Neural Network Using Behavior-Based Characteristics
Jun Sheng Hu, Zhitang Li, Zhengbing Hu, Dezhong Yao, Junfeng Yu · 2008
The paper proposes the use of the Complex-Valued Neural Network to detect spam. The main contributions of this work are two-fold. First, we present a new model based on the CVNN for classifying personal E-mails. We changed the input of the email into 2-dimensional vector. The Complex-Valued Neural Network is superior for handling 2-dimensional vector data stream, because the input of the complex-valued neural network has real part and imaginary part. Second, the Behavior-based Characteristics are extracted as most important features of the E-mail. The results reveal that the proposed technology is reliable, efficient and scalable.