Researchers at a US university have developed software which they claim can identify fake hotel reviews by analysing its linguistic structure.
The group from Cornell University in New York state claim its technology tested 800 reviews of hotels in Chicago and was able to pick out “deceptive reviews” with close to 90% accuracy.
This was achieved by asking a group of people to deliberately write fake positive reviews of properties which were later compared with an equal number of verified truthful reviews.
The university says human beings are “lousy” at identifying false reviews online (it asked a judging panel to try), but the software analyses sentiment, word structure and grammar within a review to establish whether it should be flagged up.
“Truthful hotel reviews, for example, are more likely to use concrete words relating to the hotel, like ‘bathroom’, ‘check-in’ or ‘price’.
“Deceivers write more about things that set the scene, like ‘vacation’, ‘business trip’ or ‘my husband’. Truth-tellers and deceivers also differ in the use of keywords referring to human behavior and personal life, and sometimes in features like the amount of punctuation or frequency of large words.
“In parallel with previous analysis of imaginative versus informative writing, deceivers use more verbs and truth-tellers use more nouns.”
One of the students involved in the project, Myle Ott, cautions that the work so far is only validated for hotel reviews (presumably because the technology could be applied to other verticals where reviews are used) and on the subset of properties in Chicago.
Although companies such as TripAdvisor claim to have clever software behind the scenes to help identify fake reviews on their sites, Cornell students see the technology being applied in the industry perhaps as a “first round filter”, allowing sites to score potentially false reviews which require further investigation.