“…Authors used various classi er algorithms as RF, Decision Trees (DT), Decision Table, Random Tree, KStar, Bayes Net and Simple Logistic, Content . They calculated classification, Accuracy-94.5% FPR-4.1% FNR-6.6% for spam dataset[11].Alharthi,Alhothali and Moria collected 10,000 Arabic tweets dataset for prediction.Authors used machine learning algorithms with Long Short Term Memory and word embedding feature representation.The system classification accuracy depends on tweet length and evaluated values 0.97,0.98,0.95 for Accuracy, Precision and Recall respectively[12].Liu, Pang and Wang used 97,839 Restaurant with 31,317 Hotel review dataset for classi cation. They used Machine Learning techniques, Bi-LSTM and multimodal neural network model to analyze the effective features to improve the performance, Recall-0.80 Precision-0.82 F1-score-0.81[13].Saidani, Adi and Allili organized dataset from, Enron Corpus consisting of 2,893 messages with 2,412…”