Real Time Detection of Spam Messages in Chat Systems using LSTM Networks

1 Manvanth Cooly , S.Thulasee Krishna2
1PG Student, Department of CSE, Professor, Department of CSE
Sree Rama Engineering College Tirupati,517520. AP India
c.manvanth1998@gmail.com , thulasikrishna1988@gmail.com

Abstract—The proliferation of short message service (SMS) texts is directly attributable to the proliferation of better mobile devices in recent years. Since then, spam hasbeen steadily rising, and mobile devices are the new  vector for this epidemic. Although email is still theprimary vehicle for spam, text messaging services are rapidly overtaking it. When spam messages start piling up on everyone's phone, it's annoying. Research on theproblem is ongoing, and there are several methods  availablefor detecting and decreasing spam transmissions. Classifying messages sent over the shortmessaging service (SMS) as spam is no easy feat. Several studies have used various machine learning techniques, such as SVM, Random Forest, and Naive Bayes (NB), to examine this matter. However, due to their inherent limitations, these approaches are unable to reliably classify all forms of spam. Finding a more reliable and accurate procedure requires a thoroughinvestigation. To address this issue, we introduce LongShort-Term Memory (LSTM), a cutting-edge RNNarchitecture that incorporates memory cells into itsGating Mechanism.  Artificial intelligence,LSTM, natural language processing, and machine learning are all concepts