Machine-learning for combinatorial decisioning

A method comprising:

• generating, by a processor, a matrix based on features of a plurality of reference entities and a respective allocation of value provided to each reference entity of the plurality of reference entities;
• applying, by the processor, first-stage machine-learning on the matrix to identify relevant features of the plurality of reference entities that correlate with the respective allocation of value provided to each reference entity;
• accessing, by the processor, an identity of a target entity and target features of the target entity;
• applying, by the processor, second-stage machine-learning to generate a cluster comprising the target entity and a set of the plurality of reference entities based on the relevant features of the plurality of reference entities and the target features of the target entity, the set comprising some or all of the plurality of reference entities, wherein the second-stage machine-learning is an unsupervised machine-learning;
• and
• determining, by the processor, a distribution of values allocated to the set of the plurality of reference entities in the cluster, the distribution of values being used to generate or assess a target allocation of value for the target entity.

A non-transitory computer readable medium storing instructions that when executed by a processor cause the processor to:

• generate a matrix based on features of a plurality of reference entities and a respective allocation of value provided to each reference entity of the plurality of reference entities;
• apply first-stage machine-learning on the matrix to identify relevant features of the plurality of reference entities that correlate with the respective allocation of value provided to each reference entity;
• access an identity of a target entity and target features of the target entity;
• apply second-stage machine-learning to generate a cluster comprising the target entity and a set of the plurality of reference entities based on the relevant features of the plurality of reference entities and the target features of the target entity, the set comprising some or all of the plurality of reference entities, wherein the second-stage machine-learning is an unsupervised machine-learning;
• and
• determine a distribution of values allocated to the set of the plurality of reference entities in the cluster, the distribution of values being used to generate or assess a target allocation of value for the target entity.