点击次数：

论文类型：会议论文

发表时间：2018-05-26

页面范围：76-80

摘要：The Short Message Service (SMS) has widely extended in themodern methods of communication technology. The classificationof the spam message is an interesting and prominent issue.Classifying availability of spam in SMS is a challenging task, aplenty of research has been carried out in this direction employingMachine Learning techniques such as Naive Bayes (NB), RandomForest (RF), and Support Vector Machine (SVM) for SpamClassification. Although these methods have shown adequateperformance, but are not efficient enough in terms of spamclassification. Hence, a rigorous study is needed to find a moreaccurate and robust method. To address this, we proposed a novelmethod Long Short-Term Memory (LSTMs), which is anadvanced structure of Recurrent Neural Network (RNN) that hasgating mechanism including memory cells. Additionally,Word2Vec tool has been employed in this study, which convertssimplified text into representation of words in a vector space. Toevaluate the effectiveness of our method, SMS datasets have beenused which are freely available. Experimental results prove thatproposed method outperformed state-of-the-art Machine Learningmethods like Random Forest (RF), SVM, kNN (k NearestNeighbor), Decision Tree, and providing 97.5 percent accuracy. © 2018 Association for Computing Machinery.