petrushkagoogol Posted April 27, 2018 Report Share Posted April 27, 2018 (edited) See the graphic. In this case, the algorithm related the topic to hair-dressing, which was inappropriate to say the least. I would have correlated this to "politics" based on the keywords. Assigning weightages under three major categories - Politics * nation - +1 * state - +1 * war - +1 * sociological - + 0.5 (subject category is ambiguous) Hair-dressing * (Cross)-hairs - +1 * sociological - + 0.5 Religion * Cross(-hairs) - +1 * sociological - + 0.5 You get a relative score of 3.5 / 1.5 / 1.5, which identifies the decider clearly - POLITICS How the guys at Quora arrived at their selection of title is rather obscure ?! Hmmm ........ Am I thinking straight here ? Edited April 27, 2018 by petrushkagoogol Quote Link to comment Share on other sites More sharing options...
JMJones0424 Posted April 29, 2018 Report Share Posted April 29, 2018 Yes, you're thinking straight in that the algorithm failed, but since it failed when compared to your algorithm, it is clear that your algorithm isn't a good descriptor of theirs. petrushkagoogol 1 Quote Link to comment Share on other sites More sharing options...
petrushkagoogol Posted April 29, 2018 Author Report Share Posted April 29, 2018 Yes, you're thinking straight in that the algorithm failed, but since it failed when compared to your algorithm, it is clear that your algorithm isn't a good descriptor of theirs. My algorithm is my evaluation as to how the algorithm should have been designed :vava: Quote Link to comment Share on other sites More sharing options...
DaveC426913 Posted May 29, 2018 Report Share Posted May 29, 2018 You forgot "extensions" under Hair-dressing. petrushkagoogol 1 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.