1. The document discusses methods for detecting fake online reviews using machine learning techniques. It examines using n-gram features like unigram and bigram presence and frequency, as well as review length. 2. Three classifiers - Naive Bayes, logistic regression, and support vector machines - are trained on genuine and fake reviews extracted from Yelp. The classifiers are evaluated using 5-fold cross validation. 3. Logistic regression achieved the highest accuracy of 50.6% for unigram presence, while Naive Bayes achieved the highest accuracy of 49.95% for unigram frequency. Detection accuracy was highest for linguistic features and lowest for review length alone.