Computer-implemented methods, computer-readable media, and systems for identifying causes of loss

Abstract

Another aspect of the invention provides a system for identifying causes of loss from insurance claims data comprising a plurality of unstructured or semi-structured insurance claims data. The system includes: a processor; and computer-readable memory containing instructions to: implement an interface programmed to receive insurance claims data comprising one or more insurance claims; store the insurance claims data in the computer-readable memory; and invoke execution of the method as described herein on the processor.

1 . A computer-implemented method of identifying causes of loss from insurance claims data comprising a plurality of unstructured or semi-structured insurance claims, the computer-implemented method comprising:
 loading insurance claims data comprising a plurality of unstructured or semi-structured insurance claims into memory on a computer;   for each of at least a subset of the insurance claims within the insurance claims data, creating a corresponding pre-processed claim record by:
 tokenizing the insurance claims loaded into memory to separate words in the insurance claims loaded into memory from punctuation; 
 lemmatizing the words in the insurance claims loaded into memory to map morphological variations onto a common base word; 
 removing stop words from the insurance claims loaded into memory; 
 removing punctuation and numbers from the insurance claims loaded into memory; and 
 replacing abbreviations and common typographical errors with associated words previously-defined in a data dictionary stored in memory; 
   creating a Term Frequency-Inverse Document Frequency (TF-IDF) matrix in memory detailing relative frequency of a plurality of n-word terms within at least a subject of the pre-processed claim record, wherein n is a positive integer;   selecting a plurality of features from the TF-IDF matrix; and   creating a binary classifier for each of a plurality of causes of loss.   
     
     
         2 . The computer-implemented method of  claim 1 , wherein n is an integer between 1 and 6. 
     
     
         3 . The computer-implemented method of  claim 1 , wherein the selecting step comprises applying a chi-squared test for each of the plurality of features within the TF-IDF matrix. 
     
     
         4 . The computer-implemented method of  claim 1 , further comprising:
 applying the binary classifiers against a plurality of the pre-processed claim records not previously used in creating the binary classifiers to identify one or more causes of loss.   
     
     
         5 . The computer-implemented method of  claim 4 , further comprising:
 if one or more of the insurance claims was not classified by any of the binary classifiers, designating the insurance claim as unclassifiable.   
     
     
         6 . The computer-implemented method of  claim 4 , further comprising:
 selecting a single cause of loss from the one or more causes of loss identified by the binary classifiers based upon a pre-defined hierarchy.   
     
     
         7 . The computer-implemented method of  claim 4 , wherein the binary classifiers are applied in a previously specified priority order. 
     
     
         8 . The computer-implemented method of  claim 4 , wherein the binary classifiers are applied to identify a single cause of loss. 
     
     
         9 . The computer-implemented method of  claim 4 , further comprising:
 identifying whether a plurality of the insurance claims was preventable, non-preventable, or partially preventable based on previously stored associations between causes of loss and prevention techniques.   
     
     
         10 . The computer-implemented method of  claim 9 , further comprising:
 identifying one or more prevention techniques for a plurality of the insurance claims identified as preventable or partially preventable based on previously stored associations between causes of loss and prevention techniques.   
     
     
         11 . The computer-implemented method of  claim 10 , further comprising:
 identifying which of the one or more prevention techniques are associated with a highest aggregate or average loss over the plurality of preventable or partially preventable insurance claims.   
     
     
         12 . The computer-implemented method of  claim 1 , further comprising:
 discarding terms having a total frequency over the plurality of insurance claims of less than or equal to 2 before either creating the TF-IDF matrix or selecting a plurality of features from the TF-IDF matrix.   
     
     
         13 . A computer-implemented method of identifying causes of loss from insurance claims data comprising a plurality of unstructured or semi-structured insurance claims, the computer-implemented method comprising:
 loading insurance claims data comprising a plurality of unstructured or semi-structured insurance claims into memory on a computer;   for each of at least a subset of the insurance claims within the insurance claims data, creating a corresponding pre-processed claim record by:
 tokenizing the insurance claims loaded into memory to separate words in the insurance claims loaded into memory from punctuation; 
 lemmatizing the words in the insurance claims loaded into memory to map morphological variations onto a common base word; 
 removing stop words from the insurance claims loaded into memory; 
 removing punctuation and numbers from the insurance claims loaded into memory; and 
 replacing abbreviations and common typographical errors with associated words previously-defined in a data dictionary stored in memory; 
   applying the binary classifiers created using the method of  claim 1  against a plurality of the pre-processed claim records to identify one or more causes of loss for each of the insurance claims within the insurance claims data.   
     
     
         14 . The computer-implemented method of  claim 13 , further comprising:
 if one or more of the insurance claims was not classified by any of the binary classifiers, designating the insurance claim as unclassifiable.   
     
     
         15 . The computer-implemented method of  claim 13 , further comprising:
 selecting a single cause of loss from the one or more causes of loss identified by the binary classifiers based upon a pre-defined hierarchy.   
     
     
         16 . A system for identifying causes of loss from insurance claims data comprising a plurality of unstructured or semi-structured insurance claims data, the system comprising:
 a processor; and   computer-readable memory containing instructions to:
 implement an interface programmed to receive insurance claims data comprising one or more insurance claims; 
 store the insurance claims data in the computer-readable memory; and 
 invoke execution of the method of any of  claims 1 - 15  on the processor.