LSA Based Classification of Advertising Spam Reviews

In this study, methods to classify advertising reviews from shopping mall reviews are suggested. Advertising reviews are mostly written by companies and contain advertising contents. There are a few studies regarding the classification of opinion spam documents, which is very rare in foreign studies; however, there are no studies that classify advertising reviews from Korean reviews. In this study, the Naive Bayes Classifier was used to classify review documents and the POS （Part-of-Speech）-Tagging and bigram methods were used to extract specific words. The frequency calculation methods for the probability value of specific words were： （1） The general number of appearances of words （2） the frequency calculation of specific words through the suggested Latent Semantic Analysis （LSA）, and by recalculating the result from （1） in （2）, the performances of each method were compared. As a result, the methods from （2） showed 88.43% accuracy which is 8.89% higher than 79.54% which was the previous result from using the POS-Tagging ＋ Bigram method. Therefore, it was proved that the method suggested in this study is effective at classifying or extracting advertising reviews from Korean product review documents.

Automated Spam Review Detection Using Hybrid Deep Learning on Arabic Opinions

Online reviews regarding purchasing services or products offered are the main source of users’opinions.To gain fame or profit,generally,spam reviews are written to demote or promote certain targeted products or services.This practice is called review spamming.During the last few years,various techniques have been recommended to solve the problem of spam reviews.Previous spam detection study focuses on English reviews,with a lesser interest in other languages.Spam review detection in Arabic online sources is an innovative topic despite the vast amount of data produced.Thus,this study develops an Automated Spam Review Detection using optimal Stacked Gated Recurrent Unit(SRD-OSGRU)on Arabic Opinion Text.The presented SRD-OSGRU model mainly intends to classify Arabic reviews into two classes:spam and truthful.Initially,the presented SRD-OSGRU model follows different levels of data preprocessing to convert the actual review data into a compatible format.Next,unigram and bigram feature extractors are utilized.The SGRU model is employed in this study to identify and classify Arabic spam reviews.Since the trial-and-error adjustment of hyperparameters is a tedious process,a white shark optimizer(WSO)is utilized,boosting the detection efficiency of the SGRU model.The experimental validation of the SRD-OSGRU model is assessed under two datasets,namely DOSC dataset.An extensive comparison study pointed out the enhanced performance of the SRD-OSGRU model over other recent approaches.