Abstract: Any unwanted or useless text message delivered to a mobile phone through Short Message Service (SMS) is called as a spam SMS. The spam SMS issue is gradually increasing along with the increase in the  use  of  text  messaging.  Users  usually do not like receiving such messages as they are just disturbing, and here arises the need for spam filters. The proposed system focuses on detecting the spam messages by identifying the features of each messages contained in UCI machine learning repository. A message contains different valid features, making it as either spam or ham. The proposed method makes use of three machine learning approaches called Deep Learning, Naive  Bayes, and Random Forest approach. Finally, a comparison is made among the three approaches in order to identify which technique gives the best performance in this particular task. Interestingly, each of them are so close to each other based on their performance, and gives a promising result in the task of SMS spam detection.

Keywords: SMS Spam Detection, Natural Language Pro- cessing (NLP), Deep Learning, Naive Bayes, Random Forest

Open chat
Chat with IJARCCE