In recent years, online reviews have become the most important resource of customers' opinions. These reviews are used increasingly by individuals and organizations to make a purchase and business decisions the fundamental truth of large-scale data sets is not yet available and most existing approaches to supervised learning are based on pseudo- false. Identifying these spammers and spam content is a very hot topic of research, and although a large number of studies have been conducted recently for this purpose, the methodologies presented so far have barely detected spam reviews, and no one of them shows the importance of any type of extracted feature. In this study, we propose a new framework, called NetSpam, which uses spam features to model audit datasets as heterogeneous information networks to map the spam detection procedure into a classification problem in those networks. Using the importance of spam features helps us get better results in terms of different metrics experienced in real-time data sets from Yelp and Amazon sites.