Abstract: Recommender systems have become an essential component in a wide range of web services. It is believed that recommender systems recommend a user items (e.g., products on Amazon) that match the user’s preference. Now a days, a serious a part of everyone trusts on content in social media like opinions and feedbacks of a subject or a product. The liability that anyone can begin a survey provides a brilliant chance to fake co-visitations to compose spam surveys about products and services for various interests. Recognizing these fake co-visitations and therefore the fake content may be a wildly debated issue of research and in spite of the very fact that a powerful number of studies are done as lately toward thisend, yet thus far the procedures set forth still scarcely distinguish fake reviews, fake co-visitations, and none of them demonstrate the importance of every extracted feature type. During this investigation, we propose a completely unique structure, named Fake review detection system, which uses spam highlights for demonstrating review datasets to style fake co-visitations detection method into a classification issue in such networks. Utilizing the importance of fake features help we to accumulate better outcomes regarding different metrics on review datasets. We also discuss strategies to mitigate our attacks.Keywords: Amazon, Machine learning, Product.