Meme discovery system

• US 9,967,321 B2
• Filed: 04/10/2015
• Issued: 05/08/2018
• Est. Priority Date: 08/11/2014
• Status: Active Grant

• Current Assignee: Systems & Technology Research, LLC
• Original Assignee: Systems & Technology Research, LLC
• Inventors: Sandell, Nils F., Pioch, Nicholas J., Jones, Eric K.
• Primary Examiner(s): Dalencourt, Yves
• Application Number:
  US14/683,990
• Publication Number:
  US 20160080476A1
• Time in Patent Office: 1,124 Days
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46
• CPC Class Codes: H04W 4/08 User group management
• CPC Class Codes: H04W 4/21 for social networking applications

Abstract

A system for discovering internet memes from a social media stream includes a meme detector for grouping together incoming posts that contain a common set of anomalous terms, the meme detector comparing the frequency of use of the anomalous terms against normalcy models to ensure that the use of the anomalous terms is non-recurring. The system additionally includes a meme tracker for classifying each cluster of related posts as a particular meme and, in turn, associating new related posts from the media stream with the classified cluster. In this manner, the particular set of anomalous terms used to classify each meme can be augmented to include other discriminating terms. Accordingly, the parameters, or features, associated with each meme are continuously monitored and refined, as needed, to ensure that the most comprehensive and reflective cluster of posts for each meme is collected from the social media stream in an efficient fashion.

Claims

1. A system for discovering internet memes from a social media stream generating a plurality of posts, each post comprising at least one term, the system comprising:

(a) a meme detector for grouping together a first set of posts from the social media stream, each of the posts in the first set including at least one common term, the meme detector characterizing the first set of posts as an internet meme using the at least one common term; and

(b) a meme tracker for classifying the internet meme, the meme tracker continuously monitoring posts from the social media stream and associating additional posts related to the internet meme with the first set of posts to yield a combined set of posts for the internet meme;

(c) wherein the meme detector includes a normalcy model that compiles historical data relating to the use of the at least one common term in the plurality of posts from the social media stream over a defined period of time, the historical data being stored as a normalcy table, the meme detector characterizing the first set of posts as an internet meme if use of the at least one common term exceeds use of the at least one common term in the normalcy table by a predefined threshold.

2. The system as claimed in claim 1 wherein each of the first set of posts includes one or more anomalous bigrams.

3. The system as claimed in claim 2 wherein the meme detector characterizes the first set of posts as an internet meme by creating a weighted undirected graph using the one or more anomalous bigrams from the first set of posts.

4. The system as claimed in claim 3 wherein the meme detector characterizes the first set of posts by identifying a trigram with the highest count from the weighted undirected graph.

5. The system as claimed in claim 1 wherein the meme tracker classifies the internet meme using first and second level classifiers designed for high speed and precision.

6. The system as claimed in claim 5 wherein the first level classifier filters out a majority of the plurality of posts from the media stream to yield a first subset of posts.

7. The system as claimed in claim 6 wherein the first level classifier is a decision tree classifier.

8. The system as claimed in claim 6 wherein the second level classifier compares features in the first subset of posts in order to classify the internet meme.

9. The system as claimed in claim 8 wherein the second level classifier is a Naive Bayes classifier.

10. The system as claimed in claim 1 wherein the meme tracker continuously monitors the plurality of posts from the social media stream in view of the classified internet meme.

11. The system as claimed in claim 10 wherein the meme tracker periodically associates new posts from the social media stream that relate to the classified internet meme with the first set of posts to yield the combined set of posts for the internet meme.