Spam Detection using Word Embedding-based LSTM

Abstract

Spam email is a well-known problem in network communications. A large number of spam emails in mailbox caused inconvenience and might lead to cybercrime. Effective spam mail detection is one of challenges research topics. This paper proposed a new spam mail detection approach that can be applied for sentiment analysis. A dataset of 5,572 messages was tested using Word embedding Techniques such as Bag of words, Hashing, and Long short-term memory network Algorithm (LSTM). The results showed the proposed method performance are equal 98%, 93%, 95%, and 98% interns of precision, recall, F1-score, and accuracy respectively.