For anyone who doesn’t know, every algorithm has flaws. Google Search has flaws. PeopleRank has flaws. And so does my algorithm for Quantifying Information Scarcity among the Quora users have flaws. The trick is not to give up hope and think of a new algorithm, but to instead improve the algorithm you already have.
Tony Miller suggested that I use the popularity of a topic as a factor, and I refused because the amount of views and upvotes an answer gets, is relative to the amount of people following it.
Ariel Williams had the lowest score in the 100-499 category, despite posting loads of answers in loads of different widespread categories, which further reinforces why I should still refuse.
I find 2 reasons for what made some of the people’s metrics skewed.
- Posting too many answers at once, so their followers don’t see all the answers they posted on that day.
- People who had specialist knowledge that was so specialist, that nobody in the Quora Community cared about what they had to say.
Using my human judgement, I have found what are believed to be abnormalities in the results to be coming from Adisa Nicholson, Viola Yee and Alton Sun. To rectify these abnormalities and have consistently accurate results, I will find a new metric that is affected by the original metric of distribution-of-upvotes-among-answers, and use my new 2nd metric to influence the original 1st metric, to be able to gather more accurate results, and come up with a more accurate algorithm.
Now it’s time to work out the distribution-of-upvotes-among-answers again but for the topics that Quora Users have their most answers with.
The lower your number, the better you are!
Alex Regueiro – Movies – 1.16
Alton Sun – Human Behavior – 2.16
Maria Suyay Videla – Medicine and Healthcare – 1.9
Marianne Baker – Music – 3 (has 6 answers)
Hamza Alsbaihi – Survey Question – 3
Liz Mullen – Psychology – 1.19
Tony Miller – Atheism – 1.32
Jan Leadbetter – Dating and Relationships – 1.3
Antonia Anni – Religion – 1.77
Ariel Williams – Religion – 1.25
Everyone got more upvotes than answers for the topic that they have answered in the most.
Christopher Huang – Apple (company) – 1.94
User – Philosophy – 1.09
Charles Lyell – Dopamine – 3
Marie Stein – Movies – 1.06
Elizabeth L. Mead – Movies – 1.83
David Urquhart – Quora Community – 1.12
Mark Simchock – WordPress – 2.67
Viola Yee – Food – 1.55
Bevan Audstone – History – 1.35
Lisa Galarneau – Psychology – 1.55
Alex K. Chen – Astrology – 1.15
Stephanie Vardavas – Survey Question – 1.18
Adisa Nicholson – Question That Contains Assumptions – 1.90
I think it’s very telling, the topics that people have the most answers, representing what sort of person they are.
To check whether the metrics in my previous post about the distribution-of-upvotes-among-answers, are skewed, I would have to compare my more accurate and specific 2nd metric, to my lesser accurate and more broad 1st metric, to check whether they remain at a constant increasing velocity or not.
by how the two lines on both graphs, follow each other’s path really
well, I consider that there aren’t any abnormalities within last week’s
metrics, when Quantifying Information Scarcity, as far as the Quora Communityis concerned.
It seems to be that any previously thought abnormalities in the results, come ultimately from the questions and people on Quora, not representing their specialist skillset. This is something that needs more research. Viola Yee would know.
Checking the ratio for the all answers and most answered topic metric, there are some people who broke out of the ratio for their cumulative frequency set.
For 100-499 answers, these people broke out of the common 0.9-1.11 to 1 ratio.
Maria Suyay Videla, Marianne Baker, Hamza Alsbaihi
For 500+ answers, these people broke out of the common 1.0 to 1.2 to 1 ratio.
Charles Lyell, Elizabeth L. Mead, Mark Simchock
It seems to be that the metrics are accurate in relation to the Quora Community subset of society, but that these metrics are no way an indicator of someone’s actual specialist knowledge in the outside world or for knowledge that doesn’t match the interests of the Quora Community.
If people are posting answers to questions that the Quora Community doesn’t care about, or opinionated answers that the Quora Community doesn’t agree with, then that will make the metrics not as good as they could be. Also people with more subjective answers, have higher scores than people who answer in more objective topics.
More research is needed on this, because there is no way that Viola Yee should have got the poor high score of 1.71. Also most my answers are inside Question That Contains Assumptions where I post about my subjective way of how I view the world, in fact most my answers are about how I view the world, hence my higher score.
I need to find a way to take into account the topics that people here aren’t so much interested in, and use that for metrics. Viola Yee has lots of answers to questions where she’s the only one who answered the question. I also need to find a way to take into account, the topics that are more subjective.
If you would like your metric, reply to this post. Or you can reply below about the topic you got listed for.