System for extracting customer feedback from a microblog site

• US 9,626,427 B2
• Filed: 02/11/2016
• Issued: 04/18/2017
• Est. Priority Date: 04/27/2012
• Status: Active Grant

• Current Assignee: Benbria Corporation
• Original Assignee: Benbria Corporation
• Inventors: Fraczak, Wojciech, Du, Ying
• Primary Examiner(s): Saeed, Usmaan
• Assistant Examiner(s): VO, CECILE H
• Application Number:
  US15/041,286
• Publication Number:
  US 20160162570A1
• Time in Patent Office: 432 Days
• Field of Search: 707737, 707E17046, 707E17089, 707E17109, 707E17008, 707E17044, 707E17141, 707805
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/335 Filtering based on additional data, e.g. user or group profiles filtering in web context G06F16/9535, G06F16/9536
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06Q 10/063 Operations research, analysis or management
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: H04L 43/04 Processing captured monitoring data, e.g. for logfile generation

Abstract

A system for extracting customer feedback from a microblog site includes a retrieval unit coupled to the microblog site to capture microblog updates. A filter unit coupled to the retrieval unit filters the captured microblog updates according to filter criteria that remove non-actionable items from the captured microblog updates. A learning unit coupled to the filter unit prioritizes the filtered microblog updates, and a classification unit coupled to the learning unit classifies the filtered and prioritized microblog updates. An action unit coupled to the classification unit performs appropriate actions based on the classified, filtered and prioritized microblog updates.

Claims

1. A method for extracting customer feedback from a first microblog site using a database, and one or more automated tools comprising a retrieval unit, a filter unit, a learning unit, and a classification unit, said method comprising:

capturing, by the retrieval unit, microblog updates posted by customers from the first microblog site;

filtering, by the filter unit, the captured microblog updates according to filter criteria that removes non-actionable items from the captured microblog updates; and

prioritizing, by the learning unit, the filtered microblog updates, wherein said prioritizing comprises

assigning probabilities to corresponding filtered microblog updates,

assigning priorities to said corresponding filtered microblog updates based on said assigned probabilities,

storing the corresponding filtered microblog updates in the database after assigning said priorities,

tagging the corresponding filtered microblog updates in the database based on the corresponding assigned priorities; and

classifying, by a classification unit, the filtered and prioritized microblog updates, and

wherein said classifying comprises selecting at least some of the prioritized microblog updates from the database based on said tagging.

2. The method of claim 1, wherein said assigning of probabilities further depends on influence scores of the customers determined using one or more scoring services different from the first microblog site.

3. The method of claim 1, wherein said assigning of probabilities further depends on proximities of the customers to businesses associated with the filtered microblog updates, said proximities determined by analyzing metadata, hashtags associated with said filtered microblog updates, and check-in information obtained from one or more sites different from said microblog site.

4. The method of claim 1, wherein said prioritizing comprises discarding at least one of the filtered microblog updates based on said tagging.

5. The method of claim 4, wherein said discarding is based on an intelligent dropping policy.

6. The method of claim 1, wherein said classifying further comprises determining whether each selected microblog update is actionable.

7. The method of claim 5, wherein said discarding is further based on congestion levels.

8. The method of claim 1, wherein said assigning of probabilities is based on popularity of each of the customers based on number of followers of said customer on said first microblog site.

9. The method of claim 1, wherein said assigning of probabilities is based on results from monitoring one or more microblog sites different from said first microblog site for one or more postings made by the customers.

10. The method of claim 1, further wherein each said assigned probability is used to indicate confidence that the corresponding filtered microblog update is actionable.