A Revised Converter Paradigm Designed for Spam Message Exposure

Abstract

Within this paper, we point to consider the plausibility of recognizing spams in mobile phone sms messages by recommending an improved Converter method. This method is planned for recognizing spams in SMS messages. We use “Spam Collection v.1 dataset” as well as “UtkMl's Twitter Spam Location Competition” dataset to evaluate our proposed spam Detector, with a number of well-known machine learning classifiers and cutting-edge SMS spam detection techniques serving as the benchmarks. In our paper, we use networks such by way of long short term memory (LSTM), bi-directional LSTM, and encoder-decoder LSTM models which are recurrent neural networks. Our investigations on SMS spam detection demonstrate that the proposed improved spam Converter outperforms all other alternatives regarding accuracy, F1-Score and recall. Additionally, the suggested model performs well on UtkMl's Twitter dataset, suggesting a favorable chance of applying model to other similar issues.