| CPC G06Q 30/0631 (2013.01) | 20 Claims |
|
1. A method, by at least one processor, for recommending products based on characteristics of a customer's household, the method comprising:
receiving, from a computing device, a search query submitted by the customer;
determining, from a plurality of age dependent products each having an associated developmental stage, a subset of age dependent products based on engagements by the customer's household;
retrieving, from a database, a plurality of developmental stages associated with each of the age dependent products in the subset;
determining a probability density function whose value at each data point represents a relative likelihood that the customer's household has a corresponding number of juveniles by performing a multivariate kernel density estimation upon the retrieved plurality of developmental stages;
determining a number of developmental stages associated with the customer's household based on the probability density function and a Gaussian mixture model performed upon the retrieved plurality of developmental stages;
recommending selective age dependent products of the plurality of age dependent products to the customer's household based upon the search query and the determined number of developmental stages associated with the customer's household;
transmitting, in response to the search query, the selective age dependent products to the computing device as search results; and
validating performance of the Gaussian mixture model using an evaluation metric based on statistical reasoning.
|