You are viewing a single comment's thread from:

RE: Natural Language Text Analysis of most Up voted and Down voted posts in December 2017

in #utopian-io7 years ago

I can see why "source" is the doom word, makes my veins twist in annoyance when i see it headlining the spamland... very interesting to read about, good efforts! :D

Sort:  

What is the significance of the word “source”?

I just think it's a bit of a presumptious, rather manipulative and often bias word in the context of peoples blogs. It's an individual selfishly claiming they hold all the info you need, come along and upvote me me me, when clearly it's mostly linking to shit, fake and cheap posts... it's the source of all the problems hohoho :D

I agree that most blogs, particularly the “successful” ones are mostly self righteous, puffed up, pseudoscience, pontificating tosh.

Most of the authentic stuff doesn’t see the light of day

"pontificating tosh" hahah nice

It's a shame that the only people who care and are willing to make efforts are also the people that spends hours writing the quality posts. So time consumed. Whilst the self righteous and spammers will always be there to abuse a system and quickly in numbers. Still racking brainz on ideas of how to improve... the tagging words are interesting though, giving me things to think about. I like the psychology behind it. :)

I think auto-tagging might be viable. Where an algorithm looks at word clusters and associations then assigns a tag... would prevent tag spamming

Sounds like a good idea! ... How easy would it be to implement though? I definitely think the tagging angle is one of the prime solutions... really hard to think of much else that doesn't require lots of monitoring and efforts from individuals.

Many already exist and public domain. Data scientists use them and pull them down from CRAN. The heavy math is already done, it’s an integration job... would need to be hard wired into the steemit platform though. Not something optional that users opt-in to.