Humans are unpredictable almost by definition. Love, relationships, and marriages are near impossible to predict with any accuracy -- there are just too many variables to catalog, index, and draw relationships between.
If even humans can't predict their own marriages half the time (to wit: about a 50% divorce rate in the US), how can a computer or a statistical model ever hope to compare?
Maybe these fellows really are doing something special, but they give no justification for their model here. They don't explain why they think they're on to something new or even how accurate they think their model is.
I didn't bother looking at the link, but with regard to the above comment, humans - at a scale - are predictable. We wouldn't have successful fields like psychology and economics if that weren't the case.
Humans are unpredictable almost by definition. Love, relationships, and marriages are near impossible to predict with any accuracy -- there are just too many variables to catalog, index, and draw relationships between.
If even humans can't predict their own marriages half the time (to wit: about a 50% divorce rate in the US), how can a computer or a statistical model ever hope to compare?
Maybe these fellows really are doing something special, but they give no justification for their model here. They don't explain why they think they're on to something new or even how accurate they think their model is.