April 11, 2021 Nagercoil, Tamil Nadu - Spam in Twitter is still widespread despite the serious actions taken by Twitter against it and it poses serious security threats to the legitimate users. Spammers on Twitter tend to behave like legitimate users in order to manipulate spam detection methods which are proposed against them. Traditional spam filtering methods are not suitable for Twitter since Twitter has some unique characteristics differ from other platforms.
Therefore, an approach which is specifically designed for Twitter and aware of the characteristics of Twitter is necessary to detect spam on Twitter. Twitter Spam Detector is a supervised Twitter spam detection framework which is specifically designed for Twitter and uses Naïve Bayes classifier based on Twitter-specific features in order to classify spammers from the legitimate users.
According to the evaluation result, Twitter Spam Detector’s accuracy and sensitivity are calculated as 0.943 and 0.913, respectively. The evaluation result encourages us to extend the proposed framework. As a future work, the authors would like to collect a larger dataset in order to investigate the effect of the features based on the Twitter graph on the spam detection. Best IEEE project center in nagercoil also, the framework can be extended by including soft computing techniques.
Filtering Spam Text Messages By Using Twitter-Lda Algorithm
Spam text messages is a recent issue which is suffered by the mobile users in Indonesia. In best project center Tirunelveli, this research proposes an approach to address spam text messages in Bahasa Indonesia. After data cleaning process, there are 985 text messages that eligible to proceed in the training and testing process. These text messages are divided into 860 spam and 125 ham. These text messages should be pre-processed before training and testing process are applied. The pre-processing steps are case folding, punctuation removal, tokenization, URL handling, phone number handling and stemming. After conducting text pre-processing, the implementation of Twitter-
LDA algorithm is applied in the training and testing process. In the implementation, Engineering Project center in Tirunelveli this research conducts five experiments. The experiments demonstrate that the lowest f-score is 93.33% and the highest f-score is 95.24%. Meanwhile, the average of the f-score is 94.26% and the accuracy is 96.49%. This result shows that the Twitter-LDA algorithm has a good performance to identify spam text messages in Bahasa Indonesia
Chettikulam, Nagercoil – 629002