Pseudo Base Station Spam SMS Identification Based on BiLSTM-Attention

Criminals use pseudo base station to send numerous spam SMS, which seriously damages the legitimate rights and interests of the masses. How to effectively identify spam short messages has important practical significance. This paper proposes a bidirectional long short-term memory network model that is based on attention mechanism (BiLSTM-Att) is proposed to identify pseudo base station spam messages. Short message text is input as word vectors to the bidirectional long short-term memory (BiLSTM) layer for extracting features, as a means of obtaining a final vector of features, the attention mechanism is combined with increasing weight. The Softmax classifier is used to classify the short message text feature vector. Experiments show that the pseudo base station spam SMS classification model based on the BiLSTM-Att has improved recognition accuracy compared with models of other types, and improves the effectiveness of text classification.