A simple interface for marking study data?

Before I can create a system that automatically classifies text, I need to manually classify a bunch of samples as a set for training / assessment. Is there any existing tool that will allow me to manually tag thousands of items without much trouble? And if not, what is the fastest way to hack something together?

As an example, suppose you have a bunch of Twitter posts. You would like to put them in certain buckets: happy, sad, funny, angry and spam. Some things go in several buckets. You can simply dump everything into a file and insert some tags with vi, but this is error prone and slow slowdown. More importantly, having a nice interface means you can talk to your colleagues about how to do a ton of work. The web, GUI, or console is not a big deal; just as fast and easy. Is there anything similar?

I hope so, although I can’t find anything with Google. If I need to build something, is there a good place to start? From rooting, my first impression is that Rails + jQuery + actions_as_taggable_on + jQuery Tokenizing Autocomplete looks fine, but I'm open to other things.

+3
source share
5 answers

In my case, I created something with a Ruby HighLine module for command line interfaces. It is not as interesting as the web interface, but it was easy to build and was used very quickly thanks to the single-character mode.

0
source

, Rails + jQuery + actions_as_taggable_on + jQuery Tokenizing Autocomplete, , !

+1

Amazon Mechanical Turk https://www.mturk.com/mturk/welcome . , , , .

+1

Excel ( )?

( ) , , (//...) , . , , . ( , , , ).

, .

, , . , !

+1

If you want to switch to high-tech (compared to my previous low-tech Excel answer), you can simply use Weka Tools , which "... contains tools for data preprocessing, classification, regression, clustering, association rules and visualization, as well as good suitable for developing new machine learning schemes. "

0
source

Source: https://habr.com/ru/post/1780782/


All Articles