Is that Yelp review real or fake? The article “A Framework for Fake Review Detection in Online Consumer Electronics Retailers” (Information Processing and Management 2019: 1234–1244) tested five different classification algorithms on a large corpus of Yelp reviews from New York, Los Angeles, Miami, and San Francisco whose authenticity (real or fake) was known. Since review styles differ greatly by city, the researchers used city as a blocking variable. The table below shows the F₁ score, a standard measure of classification accuracy, for each algorithm-city pairing. (F₁ scores range from 0 to 1, with higher values implying better accuracy.)

Test the null hypothesis that the five algorithms are equally accurate in classifying real and fake Yelp reviews at the .10 significance level.

Transcribed Image Text:

Algorithm Logistic regression Decision tree Random forest Gaussian Naïve Bayes AdaBoost NYC LA Miami .79 .73 .78 .81 .74 .81 .82 .78 .80 .72 .69 .71 .83 .79 .82 SF .77 .81 .80 .69 .82