An effective hybrid Cuckoo Search with Harmony search for review spam detection

In the recent years, online reviews are one of most important source of customer opinion. Nowadays consumer can gain knowledge about the products and service from online review resources, using which they can make decisions. This may lead to Opinion Spam, where spammers may manipulate and fake reviews to promote artificially or devalue the products and other services. Opinion spam detection is done by extracting meaningful features from the text, and identifying the spam reviews using machine learning techniques. This representation results in a very high dimensional feature space. These features are irrelevant, redundant, and noisy which may affect the performance of the classifier. Therefore, a good feature selection method is needed in order to speed up the processing rate, predictive accuracy. Evolutionary algorithms for feature selection can be used to handle these high-dimensional feature spaces which eliminate the noisy and irrelevant features. In this work, an effective hybrid feature selection technique using Cuckoo Search with Harmony search is proposed and Naive Bayes is used for classifying the review into spam and ham.