Compiling Documents Into A Timeline Per Event

Patent metadata

Field	Value
Publication number	US-2018067910-A1
Application number	US-201615256924-A
Country	US
Kind code	A1
Filing date	Sep 6, 2016
Priority date	Sep 6, 2016
Publication date	Mar 8, 2018
Grant date	—

Abstract

Representative embodiments disclose mechanisms to compile documents into a timeline document that tracks the evolution of a topic over time. Social media documents can be used to identify importance or popularity of linked documents (i.e., documents shared by social media in a post, tweet, etc.). A collection of social media documents is analyzed and used to identify a series of n-grams and a ranked list of linked documents. A subset of the ranked list is selected based upon similarity to the series of n-grams. The subset is then summarized and captured, along with underlying supporting data, into an entry of a timeline document. Related entries in different timeline documents can be linked to create a pivot point that allows a user to jump from one timeline to another. Timeline documents can be made available as part of a search performed by a query system.

First claim

What is claimed is:

1 . A method for capturing evolution of a topic over time as described by a set of documents, comprising:
accessing a set of first documents collected over a period of time from at least one social media service; extracting, from the set of first documents, a contextual vector comprising a set of representative n-grams, the set of representative n-grams describing aspects of the first set of documents; identifying a set of second documents linked to by the set of first documents; ranking the set of second documents according to a selected criterion; selecting a subset of the set of second documents based on similarity to at least a portion of the contextual vector; and creating an entry into a timeline document, the entry comprising one or more of: the subset of the set of second documents; a point of view associated with at least a portion of the subset of the set of second documents; a title; a description; and original documents from the first set of documents.

2 . The method of claim 1 , wherein the contextual vector comprises a topic.

3 . The method of claim 1 further comprising calculating the point of view for each document in the subset of the set of second documents.

4 . The method of claim 3 wherein calculating the point of view comprises:
identifying a sender of a document of the set of first documents;
accessing a profile associated with the sender;
identifying a point of view associated with the sender based on the profile; and
associating the point of view with at least one document of the set of second documents.

5 . The method of claim 4 , further comprising:
identifying a user that has interacted with the document of with the at least one document; identifying a point of view associated with the user; and associating the point of view with the at least one document.

6 . The method of claim 1 , further comprising:
identify an entry point in the timeline document; selecting a second entry point in a target timeline document; calculating a similarity score based on metadata associated with the entry point and metadata associated with the second entry point; and adding a link between the entry point and the second entry point when the similarity score exceeds a threshold.

7 . The method of claim 1 , wherein selecting a subset of the set of second documents based on similarity to at least a portion of the contextual vector comprises:
calculating a selection score for at least K documents of the second set of documents; and selecting as the subset of the set of second documents N documents having the highest selection scores of the second set of documents.

8 . The method of claim 1 , wherein extracting a contextual vector comprises:
clustering the set of first documents according to at least one subject matter; identifying those clusters that have a number of documents over a threshold and, for each cluster so identified:
extracting an n-gram for the cluster and storing the n-gram as part of the contextual vector;
identifying a set of documents in the identified cluster from a cluster of individuals;
clustering the set of documents to identify sub-topics within the set of documents, the clustering defining a set of sub-clusters;
identifying sub-clusters of the set of sub-clusters that have a second number of documents over a second threshold and extracting a sub-cluster n-gram for each identified sub-cluster; and
storing each sub-cluster n-gram as part of the contextual vector.

9 . A computing system comprising:
a processor and executable instructions accessible on a machine-readable medium that, when executed, cause the system to perform operations comprising:
accessing a set of first documents collected over a period of time from at least one social media service;
extracting, from the set of first documents, a contextual vector comprising a set of representative n-grams, the set of representative n-grams describing aspects of the first set of documents;
identifying a set of second documents linked to by the set of first documents;
ranking the set of second documents according to a selected criterion;
selecting the top K documents of the ranked set of second documents as a subset of the set of second documents;
for each of the top K documents, calculating a selection score based on similarity to at least a portion of the contextual vector;
selecting N documents from the top K documents based on the calculated selection score; and
creating an entry into a timeline document, the entry comprising one or more of:
at least a portion of the selected N documents;
a point of view associated with at least a portion of the selected N documents;
a title;
a description; and
original documents from the first set of documents.

10 . The system of claim 9 , wherein the contextual vector comprises a topic.

11 . The system of claim 9 , further comprising calculating the point of view for each of the selected N documents.

12 . The system of claim 11 , wherein calculating the point of view comprises:
identifying a sender of a document of the set of first documents; access a profile associated with the sender; identify a point of view associated with the sender based on the profile; and associate the point of view with at least one document of the set of second documents.

13 . The system of claim 11 , wherein calculating the point of view comprises:
identifying a user that has interacted with a document of the set of first documents; identify a point of view associated with the user, based on a profile associated with the user; and associate the point of view with at least one document of the set of second documents.

14 . The system of claim 11 , wherein calculating the point of view comprises:
analyzing content of a document of the second set of documents; and based on keywords identified from the analyzing, identifying at least one point of view to associated with the document.

15 . The system of claim 9 , further comprising:
identify an entry point in the timeline document; select a second entry point in a target timeline document; calculate a similarity score based on metadata associated with the entry point and metadata associated with the second entry point; and add a link between the entry point and the second entry point when the similarity score exceeds a threshold.

16 . The system of claim 9 , wherein extracting a contextual vector comprises:
clustering the set of first documents according to at least one subject matter; identifying those clusters that have a number of documents over a threshold and, for each cluster so identified:
extracting an n-gram for the cluster and storing the n-gram as part of the contextual vector;
identifying a set of documents in the identified cluster from a cluster of individuals;
clustering the set of documents to identify sub-topics within the set of documents, the clustering defining a set of sub-clusters;
identify sub-clusters of the set of sub-clusters that have a second number of documents over a second threshold and extracting a sub-cluster n-gram for each identified sub-cluster; and
storing each sub-cluster n-gram as part of the contextual vector.

17 . A machine-readable medium having executable instructions encoded thereon, which, when executed by at least one processor of a machine, cause the machine to perform operations comprising:
access a set of first documents collected over a period of time from at least one social media service; extract, from the set of first

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Alonso Omar
• Kandylas Vasileios
• Tremblay Serge-Eric

Classifications

• H04L67/306

  User profiles · CPC title

• G06F16/24578

  using ranking · CPC title

• G06F16/36

  Creation of semantic tools, e.g. ontology or thesauri · CPC title

• G06F16/285

  Clustering or classification · CPC title

• G06Q10/1091

  Recording time for administrative or management purposes · CPC title