bbansal / marketsenti Goto Github PK

Take a HTML document and extract useful text out of it.

This can be source specific.

problem

1. Read an article and tag it with companies it affects/talks about.
2. make a simple interface
3. Write one implementation

use wikipedia to extract keyword list

1. wikipedia page are comprehensive/complete info about any company
2. identify all clickables words in that company entry
   -- All useful/important keywords will be clickable
   -- to remove false positives (crappy keywords) remove all keywords whose pages doesnt have the actual company name into them.

void initialize()
List getValues(String storename, K key)
List putValues(String storeName, K key, List Values)
boolean delete(String storeName, K key)
Iterator<Pair<K,V>> fullScan(StoreName)

use

carrot2
lingpipe