Natural language processing based product substitution system and related methods

A product substitution server comprising:

• a processor and an associated memory configured to
• obtain descriptive purchased product data for a plurality of purchased products from at least one retailer,
• iteratively operate a natural language processing (NLP) algorithm to accept, as input thereto, the descriptive purchased product data, and generate, as output therefrom, a multi-dimensional representation of a relationship among the purchased products, and for each iteration,
• determine a product-to-product affinity based upon the multi-dimensional representation of the relationship by determining nearest neighbors in the multi-dimensional representation of the relationship and determining a distance between the nearest neighbors that is representative of the product-to-product affinity,
• for a given product, rank the nearest neighbors based upon the determined distances,
• apply, as input, the ranked nearest neighbors to a classifier,
• operate the classifier to determine whether the products represented by the ranked nearest neighbors are in a same group, the classifier cooperating with the NLP algorithm to train a machine learning model to classify the ranked nearest neighbors by group, and when so,
• determine at least one recommended substitute product for one of the plurality of purchased products based upon the determined product-to-product affinity, and communicate the at least one recommended substitute product to a shopper device associated with the given shopper, and when not so,
• obtain updated descriptive purchased product data for an updated plurality of purchased products from the at least one retailer for an operation of the classifier during a next iteration.

A method of processing a product substitution comprising:

• using a product substitution server to
• obtain descriptive purchased product data for a plurality of purchased products from at least one retailer for a given shopper,
• iteratively operate a natural language processing (NLP) algorithm to accept, as input thereto, the descriptive purchased product data, and generate, as output therefrom, a multi-dimensional representation of a relationship among the purchased products, and for each iteration,
• determine a product-to-product affinity based upon the multi-dimensional representation of the relationship by determining nearest neighbors in the multi-dimensional representation of the relationship and determining a distance between the nearest neighbors that is representative of the product-to-product affinity,
• for a given product, rank the nearest neighbors based upon the determined distances,
• apply, as input, the ranked nearest neighbors to a classifier,
• operate the classifier to determine whether the products represented by the ranked nearest neighbors are in a same group, the classifier cooperating with the NLP algorithm to train a machine learning model to classify the ranked nearest neighbors by group, and when so,
• determine at least one recommended substitute product for one of the plurality of purchased products based upon the determined product-to-product affinity, and communicate the at least one recommended substitute product to a shopper device associated with the given shopper, and when not so,
• obtain updated descriptive purchased product data for an updated plurality of purchased products from the at least one retailer for an operation of the classifier during a next iteration.

A non-transitory computer readable medium for processing a product substitution, the non-transitory computer readable medium comprising computer executable instructions that when executed by a processor of a product substitution server cause the processor to perform operations comprising:

• obtaining descriptive purchased product data for a plurality of purchased products from at least one retailer for a given shopper,
• iteratively operating a natural language processing (NLP) algorithm to accept, as input thereto, the descriptive purchased product data, and generate, as output therefrom, a multi-dimensional representation of a relationship among the purchased products, and for each iteration,
• determining a product-to-product affinity based upon the multi-dimensional representation of the relationship by determining nearest neighbors in the multi-dimensional representation of the relationship and determining a distance between the nearest neighbors that is representative of the product-to-product affinity,
• for a given product, rank the nearest neighbors based upon the determined distances,
• apply, as input, the ranked nearest neighbors to a classifier,
• operate the classifier to determine whether the products represented by the ranked nearest neighbors are in a same group, the classifier cooperating with the NLP algorithm to train a machine learning model to classify the ranked nearest neighbors by group, and when so,
• determining at least one recommended substitute product for one of the plurality of purchased products based upon the determined product-to-product affinity, and communicating the at least one recommended substitute product to a shopper device associated with the given shopper, and when not so,
• obtaining updated descriptive purchased product data for an updated plurality of purchased products from the at least one retailer for an operation of the classifier during a next iteration.