Computer-Implemented System and Method for Providing Classification Suggestions via Concept Injection

Abstract

A computer-implemented system and method for providing classification suggestions via concept injection is provided. Clusters of uncoded concepts are accessed. Each uncoded concept represents one or more documents including that concept. One or more uncoded concepts in one of the clusters is compared to a set of reference concepts. Each reference concept is associated with a classification code. One or more of the reference concepts most similar to the one or more uncoded concepts is placed into the cluster. A neighborhood is generated for one of the uncoded concepts in the cluster. The neighborhood includes at least one of the reference concepts placed in the cluster. One of the classification codes representative of the neighborhood of reference concepts is determined and provided as a suggestion for classification of the uncoded concept.

What is claimed is: 
     
         1 . A computer-implemented system for providing classification suggestions via concept injection, comprising:
 clusters of uncoded concepts, each uncoded concept representing one or more documents including that concept; and   a processor to compare one or more of the uncoded concepts in one of the clusters to a set of reference concepts, each reference concept associated with a classification code, to identify and place one or more of the reference concepts most similar to the one or more uncoded concepts into the cluster, to generate a neighborhood for one of the uncoded concepts in the cluster, the neighborhood comprising at least one of the reference concepts placed in the cluster, to determine one of the classification codes representative of the neighborhood of reference concepts, and to provide as a suggestion for classification of the uncoded concept, the classification code of the neighborhood.   
     
     
         2 . A system according to  claim 1 , wherein the processor selects the one or more uncoded concepts by determining a center of the cluster based on the one or more uncoded concepts and by identifying a sample comprising the one or more uncoded concepts. 
     
     
         3 . A system according to  claim 1 , wherein the processor displays at least a portion of the clusters along a vector based on a similarity of the uncoded concepts in those clusters. 
     
     
         4 . A system according to  claim 1 , wherein the classification suggestion satisfies a threshold confidence. 
     
     
         5 . A system according to  claim 1 , wherein the processor determines the neighborhood by determining an x-number of reference concepts and selecting the x-number of reference concepts in the cluster as the neighborhood. 
     
     
         6 . A system according to  claim 5 , wherein the processor identifies the x-number of reference concepts, comprising at least one of selecting the x-number of reference concepts nearest in distance to the uncoded concept and selecting the x-number of reference concepts nearest in distance to the uncoded concept for each different classification code associated with the reference concepts. 
     
     
         7 . A system according to  claim 1 , wherein the processor determines the classification suggestion via one of a minimum distance classification measure, minimum average distance classification measure, maximum count classification measure, and distance weighted maximum count classification measure. 
     
     
         8 . A system according to  claim 1 , wherein the processor identifies a different neighborhood for each uncoded concept in the cluster. 
     
     
         9 . A system according to  claim 1 , wherein each concept comprises a collection of nouns and noun phrases having common semantic meaning, wherein the nouns and noun phrases are extracted from the documents. 
     
     
         10 . A system according to  claim 1 , wherein one or more of the reference concepts most similar to the one or more uncoded concepts are placed in at least one other cluster of uncoded concepts: 
     
     
         11 . A computer-implemented method for providing classification suggestions via concept injection, comprising:
 accessing clusters of uncoded concepts, each uncoded concept representing one or more documents including that concept;   comparing one or more uncoded concepts in one of the clusters to a set of reference concepts, each reference concept associated with a classification code;   identifying and placing one or more of the reference concepts most similar to the one or more uncoded concepts into the cluster;   generating a neighborhood for one of the uncoded concepts in the cluster, the neighborhood comprising at least one of the reference concepts placed in the cluster;   determining one of the classification codes representative of the neighborhood of reference concepts; and   providing as a suggestion for classification of the uncoded concept, the classification code of the neighborhood.   
     
     
         12 . A method according to  claim 11 , further comprising:
 selecting the one or more uncoded concepts, comprising one or more of:
 determining a center of the cluster based on the one or more uncoded concepts; and 
 selecting a sample comprising the one or more uncoded concepts. 
   
     
     
         13 . A method according to  claim 11 , further comprising:
 displaying at least a portion of the accessed clusters along a vector based on a similarity of the uncoded concepts in those clusters.   
     
     
         14 . A method according to  claim 11 , wherein the classification suggestion satisfies a threshold confidence. 
     
     
         15 . A method according to  claim 11 , further comprising:
 determining the neighborhood, comprising:
 determining an x-number of reference concepts; and 
 selecting the x-number of reference concepts in the cluster as the neighborhood. 
   
     
     
         16 . A method according to  claim 15 , further comprising:
 identifying the x-number of reference concepts, comprising at least one of:
 selecting the x-number of reference concepts nearest in distance to the uncoded concept; and 
 selecting the x-number of reference concepts nearest in distance to the uncoded concept for each different classification code associated with the reference concepts. 
   
     
     
         17 . A method according to  claim 11 , further comprising:
 determining the classification suggestion via one of a minimum distance classification measure, minimum average distance classification measure, maximum count classification measure, and distance weighted maximum count classification measure.   
     
     
         18 . A method according to  claim 11 , further comprising:
 identifying a different neighborhood for each uncoded concept in the cluster.   
     
     
         19 . A method according to  claim 11 , wherein each concept comprises a collection of nouns and noun phrases having common semantic meaning, wherein the nouns and noun phrases are extracted from the documents. 
     
     
         20 . A method according to  claim 11 , wherein one or more of the reference concepts most similar to the one or more uncoded concepts are placed in at least one other cluster of uncoded concepts: