Back to results
US 11,144,979 B1Granted
Method, medium, and system for automatic data extraction from web pages and analysis thereof
AssigneePaypal, Inc.
InventorsOlga Sharshevsky, Yael Cohen, Elina Nemtsov
Life of this patent
Estimating the term…
Abstract
The present disclosure provides a method of automatically extracting data from web pages and analyzing the extracted data to generate an output. A plurality of web pages of a plurality of merchants is accessed. Based on the accessing of the web pages, a subset of the plurality of web pages is identified as inventory pages that contain information about products or services offered for sale. The inventory pages are electronically scanned to extract a price for each of the products or services. An output is generated that includes a listing of the products or services and prices associated with the products or services, respectively.
Claims20
A method, comprising:
- accessing, via one or more hardware processors, a plurality of web pages of a plurality of merchants;
- identifying, via the one or more hardware processors and based on the accessing, a subset of the plurality of web pages as inventory pages that contain information about products or services offered for sale, wherein the identifying is performed using a machine learning process or using a detection of recurring HyperText Markup Language (HTML) elements in the plurality of web pages;
- electronically scanning, via the one or more hardware processors, the inventory pages to extract first prices for each of the products or services, respectively;
- analyzing, after the first prices have been extracted, a Cascading Style Sheets (CSS) style of each of the inventory pages;
- accessing a plurality of predefined CSS styles that correspond to price changes;
- determining, at least in part based on comparing the analyzed CSS styles with the plurality of predefined CSS styles, that the first prices of the products or services of at least a subset of the inventory pages changed to second prices, respectively;
- and
- generating, via the one or more hardware processors, an output that includes a listing of the products or services of at least the subset of the inventory pages and the first prices and the second prices associated with the products or services, respectively.
5 more elements
A system, comprising:
- a non-transitory memory;
- and
- one or more hardware processors coupled to the non-transitory memory and configured to read instructions from the non-transitory memory to cause the system to perform operations comprising:
- accessing a plurality of web pages of a plurality of merchants, wherein at least some of the plurality of web pages are inventory pages that contain pricing information;
- determining, using machine learning or detection of recurring HyperText Markup Language (HTML) structures, which of the plurality of web pages are the inventory pages;
- extracting, based on an electronic pattern search of the inventory pages, at least the pricing information from the inventory pages;
- analyzing, after the extracting, a Cascading Style Sheets (CSS) style of each of the inventory pages;
- accessing a plurality of predefined CSS styles that correspond to price changes;
- determining, at least in part based on comparing the analyzed CSS style with the plurality of predefined CSS styles, that the pricing information of at least a subset of the inventory pages has changed;
- and
- generating an output based on the extracted pricing information.
8 more elements
A non-transitory machine-readable medium having stored thereon machine-readable instructions executable to cause a machine to perform operations comprising:
- accessing a plurality of web pages of a plurality of merchants;
- labeling, at least in part using a machine learning process, a first subset of the plurality of web pages as preliminary inventory pages;
- detecting a second subset of the preliminary inventory pages that each contains recurring HyperText Markup Language (HTML) elements;
- confirming the second subset of the preliminary inventory pages as true inventory pages;
- electronically scanning the recurring HTML elements in the true inventory pages;
- extracting, based on the electronically scanning, product descriptions and pricing data for a plurality of products on each of the true inventory pages;
- analyzing, after the extracting, a Cascading Style Sheets (CSS) style of each of the true inventory pages;
- accessing a plurality of predefined CSS styles that correspond to price changes;
- determining, at least in part based on comparing the analyzed CSS styles with the plurality of predefined CSS styles, that the pricing data of at least a subset of the true inventory pages has been updated;
- and
- generating an output that contains a listing of the plurality of products, the listing including the product descriptions and the updated pricing data.
8 more elements
Dependencies are inferred from the claim text.