Content insight system
AssigneeSprinklr, Inc.
InventorsDevarsh Sheth, Noman Ahmed Sheikh, Vasant Srinivasan, Pavitar Singh
Life of this patent
Estimating the term…
Abstract
An insight system identifies brands, subjects, attributes, and the sentiment conveyed for those attributes. The insight system pairs the attributes with the subjects and brands and generates metrics based on the sentiments associated with the attributes. The insight system may parse product webpages linked to the content for brand identifiers and associate the brand identifiers with the brands. The insight system provides more granular insight into user sentiment for different features associated with the brands.
Claims20
An apparatus including a hardware processor, the apparatus comprising:
- a set of instructions that, when executed by the hardware processor, cause the hardware processor to:
- parse non-structured data from webpages;
- identify subjects in the non-structured data;
- identify attributes in the non-structured data;
- individually identify sentiments for the attributes, respectively, wherein individually identify sentiments includes detecting sentiments in the non-structured data and individually associating the detected sentiments in the non-structured data with the identified attributes;
- identify brands in the non-structured data;
- pair the attributes with the subjects and brands using information taken from the non-structured data, wherein the pair the attributes with the subjects and brands produces structured data for the paired brands, subjects, and attributes, wherein the structured data is based on the sentiments individually identified for the attributes.
A processing system for identifying attributes of a topic, comprising:
- a processing device configured to:
- receive social media;
- identify subjects and attributes in non-structured data of the social media;
- individually identify sentiments for the attributes, respectively, wherein individually identify sentiments includes detecting sentiments in the non-structured data and individually associating the detected sentiments in the non-structured data with the identified attributes;
- identify topics in the non-structured data, the topics associated with the subjects and attributes;
- pair the attributes with the subjects and topics using information taken from the non-structured data, wherein the pair the attributes with the subjects and topics produces structured data for the paired topics, subjects, and attributes, wherein the structured data is based on the sentiments individually identified for the attributes.
An apparatus including a hardware processor, the apparatus comprising:
- a memory storing a set of instructions that, when executed by a hardware processor, cause the hardware processor to:
- identify products;
- parse non-structured data for subjects and attributes;
- individually identify sentiments in the content associated with the attributes, respectively, wherein individually identify sentiments includes detecting sentiments in the non-structured data and individually associating the detected sentiments in the non-structured data with the identified attributes;
- pair the products, subjects, and attributes using information taken from the non-structured data, wherein the pair the products, subjects, and attributes produces structured data;
- and
- generate a table based on the structured data, the table containing the pairings of the subjects and attributes with the products and the individually identified sentiments associated with the attributes, respectively;
- and
- display the table on a user interface.
Dependencies are inferred from the claim text.