Halfagain IM Forums
May 21, 2012, 18:43 *
Welcome, Guest. Please login or register.

Login with username, password and session length
News: Due to increasing spam we are forced to make this forum "Read Only". If you have an issue, most likely was discussed here. Use the Search function and try to find related discussions. If you can't find any helpful information, just email support at halfagain dot com"
 
   Home   Help Login Register  
Pages: [1]
  Print  
Author Topic: Lexical analysis  (Read 693 times)
0 Members and 2 Guests are viewing this topic.
roddik
Newbie
*

Karma: 10
Offline Offline

Posts: 4


« on: August 14, 2007, 17:59 »

hi guys! just bought content solution 2, tried to modify an article via built-in tools - khm.... 4 example, "BUT, our modern Western diets lead to something", here lead has to be synonimized with smth like "cause", instead of "prime, principal, chief" Sad So it would be great to check whether a word is a noun, adjective or verb etc. and  change it based on this infо. There are some obvious way of doing so: first, check on http://mw1.merriam-webster.com/dictionary what part of speech a word can possibly be and this resource shows also the most likely what it is (based on the order), then for example if there are 2 words and the first is a verb, than the second can be a noun, if the second of 2 is noun, than the first is probably an adjective, a verb before "to" is also probably a verb, also smth after a noun is unlikely to be another noun (as in "Western diets lead to"), also it would be mega great (but it seems too fantastic to be truth) to check plurals ("diets", crop the last "s" and then try to find a noun (cause it can only be a noun) to synonimize with). Please consider theese features, I understand that it's impossible to create a perfect tool, but adding them will take the synonimizing quality much higher and let us not worry about readability... Sincirely, roddik.

PS: there is DMI that uses only the site, and it creates such nice lists on e.g. nouns from a random text... the quality is I believe >95% (!!!!!)

PSS: shit, I've posted in a wrong section.... please transfer it to the "feature request"
« Last Edit: August 14, 2007, 18:29 by roddik » Logged
tomtom
Jr. Member
**

Karma: 10
Offline Offline

Posts: 127


« Reply #1 on: August 14, 2007, 23:47 »

Hmm that would certainly take the automation to a whole new level! Very interesting idee..

tomtom
Logged
Omar
Here to help.
Administrator
Hero Member
*****

Karma: 10001
Offline Offline

Posts: 2687


« Reply #2 on: August 17, 2007, 21:54 »

that do sounds like a good ideea, but how douable it is i can not tell you yet. Have to speak with the programmers.. from what i can see myself, may need some kind of reosurces to make all that analyze for every word, beside the fact that may not be possible to access that site to gather information about words. it may work for a user or two, but whn thousands people using CS will start rewrite and access that site, they may start deny access because will be overwhelmed with http requests.
Logged
roddik
Newbie
*

Karma: 10
Offline Offline

Posts: 4


« Reply #3 on: August 18, 2007, 07:54 »

You can make your own server - it will store all the words that were already in use by CS users and download any unknown ones from that site - it would take just 300 MB to store info about 30000 words - so an average user will make you the max of 200 mb traffic. What do you think about it?
Logged
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.8 | SMF © 2006-2008, Simple Machines LLC Valid XHTML 1.0! Valid CSS!