This paper investigates the issue of opinion spam in product reviews, which has been largely overlooked in previous research. The authors analyze 5.8 million reviews and 2.14 million reviewers from Amazon.com to show that opinion spam is widespread. They propose novel techniques to detect spam reviews, which are categorized into three types: untruthful opinions (type 1), reviews on brands only (type 2), and non-reviews (type 3). Type 1 spam includes reviews that are misleading or malicious, while type 2 and type 3 spam are easier to detect using traditional classification methods. For type 1 spam, the authors use duplicate and near-duplicate reviews to train a model, as these are almost certainly spam. They also use lift curves to evaluate the effectiveness of their model in detecting spam reviews. The results show that the model can accurately predict spam reviews, particularly those with outlier ratings or reviews from biased reviewers. The authors also analyze the impact of reviewer rank, feedback, and product sales rank on spam detection. Overall, the study highlights the importance of detecting opinion spam in product reviews to ensure the reliability of online opinions.This paper investigates the issue of opinion spam in product reviews, which has been largely overlooked in previous research. The authors analyze 5.8 million reviews and 2.14 million reviewers from Amazon.com to show that opinion spam is widespread. They propose novel techniques to detect spam reviews, which are categorized into three types: untruthful opinions (type 1), reviews on brands only (type 2), and non-reviews (type 3). Type 1 spam includes reviews that are misleading or malicious, while type 2 and type 3 spam are easier to detect using traditional classification methods. For type 1 spam, the authors use duplicate and near-duplicate reviews to train a model, as these are almost certainly spam. They also use lift curves to evaluate the effectiveness of their model in detecting spam reviews. The results show that the model can accurately predict spam reviews, particularly those with outlier ratings or reviews from biased reviewers. The authors also analyze the impact of reviewer rank, feedback, and product sales rank on spam detection. Overall, the study highlights the importance of detecting opinion spam in product reviews to ensure the reliability of online opinions.