Data relationships in a question-answering environment

Abstract

A computer-implemented method of establishing influence relationships between data in a question-answering environment is disclosed. Establishing influence relationships can include determining a set of conditions indicating a set of user statuses and analyzing, using a first natural language processing technique, a corpus of data including a set of user data. Establishing influence relationships between data can include identifying, based on analyzing the corpus of data, a set of influence factors corresponding to a subset of the set of user data and to a subset of the set of conditions. In embodiments, establishing influence relationships can include establishing, based on the set of influence factors, a set of influence relationships between the subset of the set of user data and the subset of the set of conditions.

What is claimed is: 
     
         1 . A computer-implemented method of establishing influence relationships between data in a question-answering environment, the method comprising:
 determining a set of conditions indicating a set of user statuses;   analyzing, using a first natural language processing technique, a corpus of data including a set of user data;   identifying, based on analyzing the corpus of data, a set of influence factors corresponding to a subset of the set of user data and to a subset of the set of conditions; and   establishing, based on the set of influence factors, a set of influence relationships between the subset of the set of user data and the subset of the set of conditions.   
     
     
         2 . The method of  claim 1 , further comprising:
 generating a set of explanations for the subset of the set of conditions using the set of influence relationships.   
     
     
         3 . The method of  claim 1 , wherein:
 determining the set of conditions indicating a set of user statuses includes:
 receiving a set of symptom data for a user; 
 analyzing, using a second natural processing technique, the corpus of data; and 
 determining, by analyzing the corpus of data, that the set of conditions is indicated by a subset of the set of symptom data. 
   
     
     
         4 . The method of  claim 1 , further comprising:
 determining, using the set of influence factors, a set of relationship scores for the set of influence relationships, the set of relationship scores calculated based on the set of influence factors; and   evaluating, based on the set of relationship scores, the set of influence relationships.   
     
     
         5 . The method of  claim 1 , wherein the identifying the set of influence factors corresponding to the subset of the set of user data and to the subset of the set of conditions includes:
 identifying, using a second natural language processing technique, a first characteristic of a first user datum of the subset of user data;   identifying, using a third natural language processing technique, a second characteristic relationship of a condition of the subset of conditions; and   identifying a first influence factor of the set of influence factors using the first characteristic and the second characteristic.   
     
     
         6 . The method of  claim 5 , wherein identifying the first influence factor using the first characteristic and the second characteristic includes:
 determining, based on a fourth natural language processing technique, that the first characteristic and the second characteristic are substantially similar; and   identifying the first influence factor of the set of influence factors from the first and second characteristic in response to determining that the first and second characteristic are substantially similar.   
     
     
         7 . The method of  claim 1 , wherein the set of user data is a database including one or more types of content selected from a group consisting of: economic data, geographic data, social networking data, medical data, personal data, family history, and historical user data.