Develop an application to recognize relevant entities in texts

http://www.aaai.org/ojs/index.php/aimagazine/article/view/1798/1696 

This paper deals with the semantic integration of entities in disparate texts. To recognize entities from different sources which are represented by different names but in reality point towards the same entity. 

Another approach is to match relational tuples and reconcile inconsistent data , however, this approach works well only for structured data. Hence, we are currently considering semantic integration approach and it's scope to recognize relevant entities in a text.

Another and better approach for named entity recognition is using hidden markov model, below suggested paper claims to outperform existing rule based models and other machine learning approaches.

Why machine learning model? Rule based classifier models try to fit a sentence to manually constructed finite state patterns. Simply put, to day-to-day language constructs. This method attempts to match against a sequence of words in the same way as general regular expressions matcher, which in turn make them less robust. Hence, we use machine learning approach.

http://dl.acm.org.ezproxy1.lib.asu.edu/citation.cfm?id=1073163

• Rule-Based Approaches to NER
• Statistical Approaches to NER
• CROSS-DOCUMENT COREFERENCE RESOLUTION
  AND ENTITY LINKING