Systems and methods of natural language processing and statistical analysis to identify matching categories
First Claim
1. A method, comprising:
- receiving, in a computing apparatus, a plurality of sets of itemized data, wherein each respective set of the itemized data includes a plurality of data fields containing an item description of a product purchased by a customer from a merchant;
performing, by the computing apparatus, natural language processing on item descriptions in the itemized data;
classifying, by the computing apparatus, items described by the item descriptions into item tiers based on the natural language processing, including;
determining presence of predetermined keywords in the item descriptions, andscoring the item tiers of the items based on predetermined weights of the predetermined keywords; and
generating, by the computing apparatus, profiles of merchants and profiles of customers based at least in part on the item tiers classified from the natural language processing of the itemized data, wherein the merchants and the customers are identified in the itemized data.
1 Assignment
0 Petitions
Accused Products
Abstract
Combining the natural language processing of product descriptions and statistical analysis of payment data to classify consumers based on products purchased and merchants based on products sold. Systems and methods use natural language processing techniques to interpret the descriptions of item level purchase data to classify products that have been purchased by customers into micro-categories. Statistical deviation methods are applied to the payment data to calculate normalized mean product cost, after removing outliers. After determining the micro-categories of the products purchased and the mean product cost of the purchased products, the system and methods classify consumers and merchants into categories based at least in part on the product micro-categories, mean costs, and relative volume of product types sold by merchants to predict which consumers are likely to purchase from which merchants.
-
Citations
20 Claims
-
1. A method, comprising:
-
receiving, in a computing apparatus, a plurality of sets of itemized data, wherein each respective set of the itemized data includes a plurality of data fields containing an item description of a product purchased by a customer from a merchant; performing, by the computing apparatus, natural language processing on item descriptions in the itemized data; classifying, by the computing apparatus, items described by the item descriptions into item tiers based on the natural language processing, including; determining presence of predetermined keywords in the item descriptions, and scoring the item tiers of the items based on predetermined weights of the predetermined keywords; and generating, by the computing apparatus, profiles of merchants and profiles of customers based at least in part on the item tiers classified from the natural language processing of the itemized data, wherein the merchants and the customers are identified in the itemized data. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18)
-
-
19. A computing apparatus, comprising:
-
a transaction handler configured to processing payment transactions in an electronic payment processing network; a portal coupled with the transaction handler and configured to communicate with transaction terminals that initiate the payment transactions in the electronic payment processing network, wherein the portal is configured to receive, using a communication channel outside the electronic payment processing network, itemized data of items purchased via the payment transactions; a data warehouse configured to store the itemized data and transaction data recording the payment transactions processed by the transaction handler; a natural language processing engine configured to performing natural language processing on item descriptions in the itemized data, classify the items described by the item descriptions into item tiers based on the natural language processing by; determining presence of predetermined keywords in the item descriptions, and scoring the item tiers of the items based on predetermined weights of the predetermined keywords; and a statistical analyzer configured to identify price tiers, volume tiers, wherein profiles of merchants and customers involved in the payment transactions are generated based at least in part on the item tiers, the price tiers, and the volume tiers.
-
-
20. A non-transitory computer storage medium storing instructions which when executed on a computing apparatus, cause the computing apparatus to perform a method, the method comprising:
-
receiving, in the computing apparatus, a plurality of sets of itemized data, wherein each respective set of the itemized data includes a plurality of data fields containing an item description of a product purchased by a customer from a merchant; performing, by the computing apparatus, natural language processing on item descriptions in the itemized data; classifying, by the computing apparatus, items described by the item descriptions into item tiers based on the natural language processing, including; determining presence of predetermined keywords in the item descriptions, and scoring the item tiers of the items based on predetermined weights of the predetermined keywords; and generating, by the computing apparatus, profiles of merchants and profiles of customers based at least in part on the item tiers classified from the natural language processing of the itemized data, wherein the merchants and the customers are identified in the itemized data.
-
Specification