4 ms·
>" knowing that three separate people are in the same coffeeshop together on several separate occasions at different times leads to the conclusion that they're
by bogomipz 9y ago
>" knowing that three separate people are in the same coffeeshop together on several separate occasions at different times leads to the conclusion that they're likely to know each other."
Does it? Three people who visit the same coffeeshop at the same time in the morning also leads to the conclusion that they all live on the same block and work 9 to 5 jobs.
- onion2k 9y agoYou missed two points. Firstly "at different times" meaning people who are all there at 3pm one day, 10am a different day, and 4:30pm on another day. That's a strong indicator that they're not there coincidently if they don't appear there at other times as well. Secondly you missed where I said "likely to know each other". Mining large data sets for relationships is all about probability, as nothing is really guaranteed to be 100% certain, but you can get to quite high confidence levels with relatively small amounts of data.
- conductr 9y agoSo you're saying the inferred information doesn't exist because of some false positives? I'd say false positives can be more dangerous because humans will have propensity to trust the algorithm and interpret it as truth
- bogomipz 9y ago>"So you're saying the inferred information doesn't exist because of some false positives?" No, exactly nowhere in my comment did I say "inferred information doesn't exist." Nor did I imply that.
- squarefoot 9y agoLots of data at hand means also one can more easily filter out false positives. 3 people in the same coffee shop can be totally strangers to each other, but if their phones start connecting to the same towers in different places all far away the coffee shop at the same times, one could conclude they're likely related someway.