“‘Computer, on the basis of likes character better predict than family'”

Researchers from the universities of Cambridge and Stanford enable computers on the basis of Facebook-likes the character of a person better able to estimate than friends and family. At 300 likes, a computer can the personality traits even better ” see ” than a partner.

For the research had 86.220 Facebook users volunteered, writes The Telegraph. An average user had 227 likes on the social networking site posted. To describe the character of a person, a number of personality traits, such as confidence, accuracy, and openness, then an algorithm, the likes of the Facebook users began to analyze. At the same time, had the participants in the study questionnaires. These lists were compared with the findings of the computer.

According to the researchers, the developed algorithm the personality traits soon to better describe if a person has a lot of likes on Facebook has posted. So would the computer a person better able to assess than a friend after 70 likes, while family members at 150 drilled likes be crossed. After watching 300 likes would be the system of the personality traits of someone even better estimate than the partner of a person.

The accuracy of the analysis software and the fact that Facebook-likes tell a lot about the character of a person by the researchers on the one hand, worrisome called, particularly in the area of privacy. They advocate for stricter laws and supervision on companies with such data collection. On the other hand, they also see advantages: such an algorithm could be used, for example, to search for a suitable job or a partner.

Although the scientists, a significant difference was found between computers and people was the absolute difference between the two groups is small. Therefore, it is unclear whether computers in the practice also proven to perform better than persons. Also found that, despite the better average performance, computer models are not the whole picture of someone’s personality could get. The average number of likes was the model on average slightly more than 30 percent of the personality to explain.