Ensemble classification algorithms having subclass resolution

Abstract

Ensemble classification algorithms having subclass resolution are disclosed. An example disclosed apparatus includes a fingerprint generator to generate a fingerprint of class probabilities of each of a plurality of samples, a distribution creator to create a distribution of the samples based on the generated fingerprints, and a distribution applicator to apply the distribution to a population to predict sub-class probabilities of each of the population.

What is claimed is: 
     
         1 . An apparatus, comprising:
 at least one memory;   instructions; and   processor circuitry to execute the instructions to:
 generate fingerprints of primary class probabilities of known samples, the known samples associated with survey data; 
 remove a first portion of the known samples in which a predicted primary class is different from a known primary class to define a second portion of the known samples in which a predicted primary class is identical to a known primary class; 
 create a distribution of predicted subclass probabilities of the second portion of the known samples, the distribution of predicted subclass probabilities based on the generated fingerprints; and 
 apply the distribution of the predicted subclass probabilities to unknown samples to determine a sub-class of ones of the unknown samples. 
   
     
     
         2 . The apparatus of  claim 1 , wherein the processor circuitry is to execute the instructions to generate the fingerprints based on class numbers that are ranked according to the respective primary class probabilities. 
     
     
         3 . The apparatus as defined in  claim 2 , wherein the processor circuitry is to execute the instructions to generate the fingerprints based on a configurable parameter that can reduce at least one of a statistical bias or variance of the fingerprint. 
     
     
         4 . The apparatus as defined in  claim 1 , wherein ones of the fingerprints represent a characteristic, category, interest, or affiliation of a respective one of the known samples. 
     
     
         5 . The apparatus as defined in  claim 1 , wherein the processor circuitry is to execute the instructions to remove a third portion of the known samples from the second portion of the known samples based on at least one of a reliability, a relevance, or a significance of the third portion of the known samples. 
     
     
         6 . The apparatus as defined in  claim 1 , wherein the processor circuitry is execute the instructions to create the distribution by pivoting a table based on at least one fingerprint value. 
     
     
         7 . The apparatus as defined in  claim 1 , wherein the processor circuitry is to execute the instructions to generate respective ones of the fingerprints by arranging the known samples in an array. 
     
     
         8 . The apparatus as defined in  claim 7 , wherein the processor circuitry is to execute the instructions to pivot the array based on attributes of the known samples. 
     
     
         9 . The apparatus as defined in  claim 7 , wherein the processor circuitry is to execute the instructions to generate the distribution by weighting at least one of the known samples. 
     
     
         10 . A method comprising:
 retrieving, by executing instructions with at least one processor, known samples associated with survey data from a database;   generating, by executing instructions with the at least one processor, fingerprints of primary class probabilities of the known samples;   removing, by executing instructions with the at least one processor, a first portion of the known samples in which a predicted primary class is different from a known primary class to define a second portion of the known samples in which a predicted primary class is identical to a known primary class;   creating, by executing instructions with the at least one processor, a distribution of predicted subclass probabilities of the second portion of the known samples, the distribution of predicted subclass probabilities based on the generated fingerprints; and   applying, by executing instructions with the at least one processor, the distribution of the predicted subclass probabilities to unknown samples to determine a sub-class of ones of the unknown samples.   
     
     
         11 . The method as defined in  claim 10 , wherein the generating of the fingerprints is based on class numbers that are ranked according to the respective primary class probabilities. 
     
     
         12 . The method as defined in  claim 11 , wherein the generating of the fingerprints is based on a configurable parameter that can reduce at least one of a statistical bias or variance of the fingerprint. 
     
     
         13 . The method as defined in  claim 10 , wherein ones of the fingerprints represent a characteristic, category, interest, or affiliation of a respective one of the known samples. 
     
     
         14 . The method as defined in  claim 10 , wherein the creating of the distribution includes pivoting a table based on at least one fingerprint. 
     
     
         15 . The method as defined in  claim 10 , wherein the generating of the fingerprints includes arranging the known samples in an array. 
     
     
         16 . The method as defined in  claim 15 , wherein the creating of the distribution of the subclass probabilities of the known samples includes weighting, by executing an instruction with the processor, at least one of the known samples. 
     
     
         17 . A non-transitory machine readable medium comprising instructions, which when executed, cause at least one processor to at least:
 access known samples associated with survey data from a database via a network;   generate fingerprints of primary class probabilities of the known samples;   remove a first portion of the known samples in which a predicted primary class is different from a known primary class to define a second portion of the known samples in which a predicted primary class is identical to a known primary class;   create a distribution of predicted subclass probabilities of the second portion of the known samples, the distribution of predicted subclass probabilities based on the generated fingerprints; and   apply the distribution of the predicted subclass probabilities to unknown samples to determine a sub-class of ones of the unknown samples.   
     
     
         18 . The machine readable medium as defined in  claim 17 , wherein the instructions, when executed, cause the at least one processor to generate the fingerprints based on class numbers that are ranked according to respective primary class probabilities. 
     
     
         19 . The machine readable medium as defined in  claim 17 , wherein the instructions, when executed, cause the at least one processor to generate the fingerprints based on a configurable parameter that can reduce at least one of a statistical bias or variance of the fingerprint. 
     
     
         20 . The machine readable medium as defined in  claim 17 , wherein ones of the fingerprints represent a characteristic, category, interest, or affiliation of a respective one of the known samples. 
     
     
         21 . The machine readable medium as defined in  claim 17 , wherein the instructions, when executed, cause the at least one processor to create the distribution by pivoting a table based on at least one fingerprint. 
     
     
         22 . The machine readable medium as defined in  claim 17 , wherein the instructions, when executed, cause the at least one processor to generate the fingerprints by arranging the known samples in an array. 
     
     
         23 . The machine readable medium as defined in  claim 22 , wherein the instructions, when executed, cause the at least one processor to create the distribution based on weighted samples of the known samples.