Learning system for pangenetic-based recommendations

1. A computer-implemented method comprising:

obtaining a set of pangenetic attributes of a set of individuals, wherein the pangenetic attributes are respectively associated with products via a set of statistical associations;

based on the set of statistical associations, determining product recommendations for a second set of individuals;

receiving, from the second set of individuals, a plurality of measures of satisfaction with the product recommendations;

learning an association of the statistical associations between (i) a set of core attributes comprising a subset of the pangenetic attributes, and (ii) a particular product of the products, wherein the set of core attributes is associated with the plurality of measures of satisfaction and the particular product, and wherein the core set of attributes is formed by removing, from the set of pangenetic attributes, at least some of the pangenetic attributes that contribute less to a statistical significance of the learned association than other of the pangenetic attributes;

receiving a representation of the subset of the pangenetic attributes that is associated with a particular individual;

receiving a request for a product recommendation for the particular individual; and

in response to the subset of the pangenetic attributes being associated with the particular individual and the request for the product recommendation, (i) based on the measures of satisfaction and the learned association, identifying the particular individual and a second individual of the second set of individuals as a pangenetic cluster, (ii) predicting a preference of the particular individual for the particular product based on a measure of satisfaction that the second individual has related to the particular product, and (iii) providing a recommendation of the particular product.

2. The computer-implemented method of claim 1 , further comprising:

obtaining a set of non-pangenetic attributes of the set of individuals, wherein the set of core attributes further comprises a subset of the non-pangenetic attributes.

3. The computer-implemented method of claim 1 , wherein the subset of the pangenetic attributes is one or more single nucleotide polymorphisms (SNPs).

4. The computer-implemented method of claim 1 , wherein the subset of the pangenetic attributes is associated with a phenotype, and wherein one or more of the measures of satisfaction indicate that the particular product is favorable to individuals exhibiting the phenotype.

5. The computer-implemented method of claim 1 , wherein the measures of satisfaction are stored in a two-dimensional item feedback matrix that maps instances of the measures of satisfaction from specific individuals to specific products.

6. The computer-implemented method of claim 1 , wherein one or more of the set of pangenetic attributes are stored in a masked fashion to prevent access by unauthorized parties.

7. The computer-implemented method of claim 1 , wherein removing, from the set of pangenetic attributes, at least some of the pangenetic attributes that contribute less to the statistical significance of the association than other of the pangenetic attributes comprises:

forming the set of core attributes by removing, from the set of pangenetic attributes, one or more pangenetic attributes for which both satisfaction and dissatisfaction with the particular product are recorded in the plurality of measures of satisfaction.

8. The computer-implemented method of claim 1 , wherein removing, from the set of pangenetic attributes, at least some of the pangenetic attributes that contribute less to the statistical significance of the association than other of the pangenetic attributes comprises:

selecting, from the second set of individuals, a subset of individuals with pangenetic attributes that have at least a threshold level of association with the plurality of measures of satisfaction; and

identifying, as the set of core attributes, a largest pangenetic attribute combination shared in common among the pangenetic attributes of the subset of individuals.

9. The computer-implemented method of claim 1 , wherein removing, from the set of pangenetic attributes, at least some of the pangenetic attributes that contribute less to the statistical significance of the association than other of the pangenetic attributes comprises:

evaluating specific pangenetic attributes by removing the specific pangenetic attributes one-by-one from the set of pangenetic attributes in order to identify contributions of each pangenetic attribute of the specific pangenetic attributes toward respective strengths of the association with the plurality of measures of satisfaction; and

forming the set of core attributes by removing, from the set of pangenetic attributes, at least one of the specific pangenetic attributes with a lower strength of association than the set of pangenetic attributes.

10. A non-transitory computer-readable medium containing program instructions that, when executed by a processor of a computing device, cause the computing device to perform operations comprising:

obtaining a set of pangenetic attributes of a set of individuals, wherein the pangenetic attributes are respectively associated with products via a set of statistical associations;

based on the set of statistical associations, determining product recommendations for a second set of individuals;

receiving, from the second set of individuals, a plurality of measures of satisfaction with the product recommendations;

learning an association of the statistical associations between (i) a set of core attributes comprising a subset of the pangenetic attributes, and (ii) a particular product of the products, wherein the set of core attributes is associated with the plurality of measures of satisfaction and the particular product, and wherein the core set of attributes is formed by removing, from the set of pangenetic attributes, at least some of the pangenetic attributes that contribute less to a statistical significance of the association than other of the pangenetic attributes;

receiving a representation of the subset of the pangenetic attributes that is associated with a particular individual;

receiving a request for a product recommendation for the particular individual; and

in response to the subset of the pangenetic attributes being associated with the particular individual and the request for the product recommendation, (i) based on the measures of satisfaction and the learned association, identifying the particular individual and a second individual of the second set of individuals as a pangenetic cluster, (ii) predicting a preference of the particular individual for the particular product based on a measure of satisfaction that the second individual has related to the particular product, and (iii) providing a recommendation of the particular product.

11. The non-transitory computer-readable medium of claim 10 , wherein the operations further comprise:

obtaining a set of non-pangenetic attributes of the set of individuals, wherein the set of core attributes further comprises a subset of the non-pangenetic attributes.

12. The non-transitory computer-readable medium of claim 10 , wherein the subset of the pangenetic attributes is associated with a phenotype, and wherein one or more of the measures of satisfaction indicate that the particular product is favorable to individuals exhibiting the phenotype.

13. The non-transitory computer-readable medium of claim 10 , wherein the measures of satisfaction are stored in a two-dimensional item feedback matrix that maps instances of the measures of satisfaction from specific individuals to specific products.

14. The non-transitory computer-readable medium of claim 10 , wherein one or more of the set of pangenetic attributes are stored in a masked fashion to prevent access by unauthorized parties.

15. The non-transitory computer-readable medium of claim 10 , wherein removing, from the set of pangenetic attributes, at least some of the pangenetic attributes that contribute less to the statistical significance of the association than other of the pangenetic attributes comprises:

forming the set of core attributes by removing, from the set of pangenetic attributes, one or more pangenetic attributes for which both satisfaction and dissatisfaction with the particular product are recorded in the plurality of measures of satisfaction.

16. The non-transitory computer-readable medium of claim 10 , wherein removing, from the set of pangenetic attributes, at least some of the pangenetic attributes that contribute less to the statistical significance of the association than other of the pangenetic attributes comprises:

selecting, from the second set of individuals, a subset of individuals with pangenetic attributes that have at least a threshold level of association with the plurality of measures of satisfaction; and

identifying, as the set of core attributes, a largest pangenetic attribute combination shared in common among the pangenetic attributes of the subset of individuals.

17. The non-transitory computer-readable medium of claim 10 , wherein removing, from the set of pangenetic attributes, at least some of the pangenetic attributes that contribute less to the statistical significance of the association than other of the pangenetic attributes comprises:

evaluating specific pangenetic attributes by removing the specific pangenetic attributes one-by-one from the set of pangenetic attributes in order to identify contributions of each pangenetic attribute of the specific pangenetic attributes toward respective strengths of the association with the plurality of measures of satisfaction; and

forming the set of core attributes by removing, from the set of pangenetic attributes, at least one of the specific pangenetic attributes with a lower strength of association than the set of pangenetic attributes.

18. A computing device comprising:

a processor;

memory; and

program instructions, stored in the memory, that upon execution by the processor cause the computing device to perform operations comprising:

obtaining a set of pangenetic attributes of a set of individuals, wherein the pangenetic attributes are respectively associated with products via a set of statistical associations;

based on the set of statistical associations, determining product recommendations for a second set of individuals;

receiving, from the second set of individuals, a plurality of measures of satisfaction with the product recommendations;

learning an association of the statistical associations between (i) a set of core attributes comprising a subset of the pangenetic attributes, and (ii) a particular product of the products, wherein the set of core attributes is associated with the plurality of measures of satisfaction and the particular product, and wherein the core set of attributes is formed by removing, from the set of pangenetic attributes, at least some of the pangenetic attributes that contribute less to a statistical significance of the association than other of the pangenetic attributes;

receiving a representation of the subset of the pangenetic attributes that is associated with a particular individual;

receiving a request for a product recommendation for the particular individual; and

in response to the subset of the pangenetic attributes being associated with the particular individual and the request for the product recommendation, (i) based on the measures of satisfaction and the learned association, identifying the particular individual and a second individual of the second set of individuals as a pangenetic cluster, (ii) predicting a preference of the particular individual for the particular product based on a measure of satisfaction that the second individual has related to the particular product, and (iii) providing a recommendation of the particular product.

19. The computing device of claim 18 , wherein the subset of the pangenetic attributes is associated with a phenotype, and wherein one or more of the measures of satisfaction indicate that the particular product is favorable to individuals exhibiting the phenotype.

20. The computing device of claim 18 , wherein the measures of satisfaction are stored in a two-dimensional item feedback matrix that maps instances of the measures of satisfaction from specific individuals to specific products.