SteemPlus 2.8 : Steem Sincerity

Steemplus is a browser extension that adds plenty of features to your Steemit (mostly) , Busy and Utopian experience. You can find the complete list of features and installation procedure in our documentation.

New Feature : Steem Sincerity

Last week we came across @andybet's great initiative : Steem Sincerity Anti-Spam API and got in touch with him in order to make it even better.

The API classifies users in 3 different categories : Human, Bot and Spammers, based on a machine learning algorithm. While I perform quite well already, it was trained on a very small data set, which can cause some classification errors.

To overly simplify. the machine learning trains itself on a few accounts that have been manually attributed to one of these 3 categories. Based on similarities it will then attribute a likelyhood for other accounts to pertain to these categories. The biggest the training set, the higher the precision will be. That's where we need you!

Read on to see how all this has been implented into SteemPlus:

For every user you'll now have a label mentioning which 'category' of user it is:

This classification is available from within the posts, the comments, on the Trending etc. lists and the feed.

Here we can see that Steemplus is considered as a human. Great I've passed the Turing test ;p

But as I said earlier, it is not perfect yet and we can see some wrong category attributions:

For example, this user is considered at a bot although I know perfectly that he is human. Or did he fool me? In his case, the fact that he is a Utopian moderator make him post often the same kind of comments, hence the error.

Same for this one who's clearly supposed to be a spammer but considered as a human.

That's the reason why we need you feedback. To give it is very simple, simply click on the category attributed to open its popup. From there you can "vote" for the category that should be attributed. @andybets' algorithm will take this into account after filtering fraudulent votes, in order to enhance the classification algorithm. This means that you will not see the effect of your vote instantly.

A simple click on one of those button will let you give your feedback and help improve the machine learning algorithm.

This algorithm is based on what you post, what you like, what you comment on Steemit. Therefore some users, the new ones for example or the ones who are never posting won't be in any category.

The main difficulty we encountered was to attribute the category to everyone in the different Steemit pages. This took time since we had to scrap a lot of different classes. The button is available

• On post's headers
  

• For the Comments
  

• On post footers
  

• On the Blog

• On the Feed / Trending / Hot / ...

Please help us by giving feedback, together we can make Steem Sincerity API an awesome tool that will help filtering spammers and bots on demand. Do not hesitate to vote for the right category the user should be in, whether to correct the algorithm or show it that it was right.

We count on you!

Coming soon

• Offline mode : Currently to use Steemplus you need to be logged via SteemConnect, but we actually don't need to be connected for all the features.

Join us on Discord

Find @andybets announcement in here.

Big shoutout to @jerrybanfield and his budgets project that allow us the speed up the development of SteemPlus!

@stoodkev and @cedricguillas for @steem-plus

Powered by @adsactly

Powered by @utopian-io

Like what I do? Vote @stoodkev for witness!

---

Posted on Utopian.io - Rewarding Open Source Contributors

---