After examining a huge dataset on human relationships in what is considered the largest laboratory on the planet, Facebook, it was discovered that, if you are in a relationship, it is best not to have many friends in common .
To target the data, the researchers looked at users who were in "a relationship," based on their Facebook profile status. Then they looked at how long it took each to go from being "in a relationship" to "single."
The study , conducted by Lars Backstrom and Jon Kleinberg , challenges our intuition. Not surprisingly, to carry out this work an experimental algorithm was used that analyzes his personal network of friends, seeking to identify his strongest relationships.
Developed in conjunction with his former Cornell thesis advisor Jon Kleinberg, Backstrom’s algorithm is sophisticated enough to independently identify your spouse or romantic partner and even predict when a breakup will occur.
The data showed the following correlation: Having a group of mutual friends was a strong predictor that the relationship would not last long. The reason is unknown, but it may have to do with meeting often with your partner and the same people end up eroding the relationship .
Backstrom’s research is part of a growing movement in companies and universities to use machine learning and vast amounts of online data to better understand human behavior and interactions and interests. This type of data analysis is revolutionizing the social sciences and changing our deep understanding of people as social beings .