Method, system and computer program product for profiling entities
View Patent ↗The present invention provides a method, system and computer program product for profiling an entity based on information obtained form at least one information source. Various contexts associated with the entity are identified. This can be achieved by using a clustering algorithm, an ontology, a thesaurus, association rules or manually by an expert. After the classified into various sets and ranked using a ranking algorithm. Thereafter, certain top ranked concepts are presented to a user as the profile of the entity.
1. A method of profiling an entity, said method comprising:
retrieving information from at least one information source using said entity as a search criteria;
clustering the retrieved information to identify contexts related to said entity, wherein said contexts are represented by a set of documents;
retrieving information corresponding to each identified context from at least one information source;
selecting features from the information retrieved at both of the retrieving steps to identify concepts associated with said entity within each identified context;
structuring the identified concepts within said each identified context by classifying said identified concepts into exactly four classification sets;
ranking said identified concepts within each set; and
profiling said entity by presenting the top ranked concepts within each set to a user,
wherein the contexts are identified by finding a set of the words or phases that occur frequently with said entity and that mutually do not appear together in documents in the information source, and
wherein said contexts are identified by finding prominent nodes, that comprise said entity, in an ontology or a taxonomy.
2. The method of claim 1 , wherein said contexts are identified by using at least one of synonyms, hypernyms, hyponyms, and meronyms of said entity found in a thesaurus.
3. The method of claim 1 , wherein said four classification sets comprise:
a set of concepts that are exclusive to said entity and unrelated to said identified context;
a set of concepts that are common to both said entity and said identified context and more frequently associated with said entity;
a set of concepts that are common to both said entity and said identified context and more frequently associated with said identified context; and
a set of concepts that are exclusive to said identified context and unrelated to said entity.
4. A method of profiling an entity, said method comprising:
identifying contexts associated with said entity, wherein said contexts are represented by a set of documents;
retrieving information corresponding to each identified context from at least one information source;
selecting features from the retrieved information to identify concepts associated with said entity within said each identified context;
structuring the identified concepts within said each identified context by classifying said identified concepts into exactly four classification sets;
ranking said identified concepts within each set; and
profiling said entity by presenting the top ranked concepts within each set to a user,
wherein the contexts are identified by finding a set of the words or phrases that occur frequently with said entity and that mutually do not appear together in documents in the information source, and
wherein said contexts are identified by using at least one of synonyms, hypernyms, hyponyms, and meronyms of said entity found in a thesaurus.
5. The method of claim 4 , said contexts are identified by finding prominent nodes, that comprise said entity, in an ontology or a taxonomy.
6. The method of claim 4 , wherein said four classification sets comprise:
a set of concepts that are exclusive to said entity and unrelated to said identified context;
a set of concepts that are common to both said entity and said identified context and more frequently associated with said entity;
a set of concepts that are common to both said entity and said identified context and more frequently associated with said identified context; and
a set of concepts that are exclusive to said identified context and unrelated to said entity.
7. A program storage device readable by computer, tangibly embodying a program of instructions executable by the computer to perform a method of profiling an entity, said method comprising:
retrieving information from at least one information source using said entity as a search criteria;
clustering the retrieved information to identify contexts related to said entity, wherein said contexts are represented by a set of documents;
retrieving information corresponding to each identified context from at least one information source;
selecting features front the information retrieved at both of the retrieving steps to identify concepts associated with said entity within each identified context;
structuring the identified concepts within said each identified context by classifying said identified concepts into exactly four classification sets;
ranking said identified concepts within each set; and
profiling said entity by presenting the top ranked concepts within each set to a user,
wherein the contexts are identified by finding a set of the words or phrases that occur frequently with said entity and that mutually do not appear together in documents in the information source, and
wherein said contexts are identified by finding prominent nodes, that comprise said entity, in an ontology or a taxonomy.
8. The program storage device of claim 7 , wherein said contexts are identified by using at least one of synonyms, hypernyms, hyponyms, and meronyms of said entity found in a thesaurus.
9. The program storage device of claim 7 , wherein in said method, said four classification sets comprise:
a set of concepts that are exclusive to said entity and unrelated to said identified context;
a set of concepts that are common to both said entity and said identified context and more frequently associated with said entity;
a set of concepts that are common to both said entity and said identified context and more frequently associated with said identified context; and
a set of concepts that are exclusive to said identified context and unrelated to said entity.
10. A system for profiling an entity, said system comprising:
means for retrieving information from at least one information source using said entity as a search criteria;
means for clustering the retrieved information into entity-context pairs in order to identify contexts related to said entity, wherein said contexts are represented by a set of documents;
means for retrieving information corresponding to each identified context from at least one information source;
means for selecting features from the retrieved information in order to identity concepts associated with said entity within said each identified context;
means for structuring the identified concepts within said each identified context by classifying said identified concepts into exactly four classification sets;
means for ranking said identified concepts within each set; and
means for profiling said entity by presenting the top ranked concepts within each set to a user,
wherein the contexts are identified by finding a set of the words or phrases that occur frequently with said entity and that mutually do not appear together in documents in the information source, and
wherein said contexts are identified by finding prominent nodes, that comprise said entity, in an ontology or a taxonomy.
11. The system of claim 10 , wherein said means for structuring the identified concepts within said each identified context comprises:
a classifier operable for classifying said identified concepts into sets with respect to each entity-context pair; and
a component operable for ranking said identified concepts within each set.
12. The system of claim 10 , wherein said four classification sets comprise:
a set of concepts that are exclusive to said entity and unrelated to said identified context;
a set of concepts that are common to both said entity and said identified context and more frequently associated with said entity;
a set of concepts that are common to both said entity and said identified context and more frequently associated with said identified context; and
a set of concepts that are exclusive to said identified context and unrelated to said entity.