Determining whether a user in a social network is an authority on a topic

Patent metadata

Field	Value
Publication number	US-10963467-B1
Application number	US-201916544799-A
Country	US
Kind code	B1
Filing date	Aug 19, 2019
Priority date	Aug 26, 2014
Publication date	Mar 30, 2021
Grant date	Mar 30, 2021

Abstract

A method involving obtaining a first plurality of topic groups (TGs), each having a membership of accounts, identifying a first plurality of accounts as authorities for an expertise topic, obtaining a second plurality of TGs with a number of accounts as members, wherein the first plurality of TGs comprises the second plurality of TGs, identifying a first frequent account which is a member in at least one of the second plurality of TGs, adding the first frequent account to the authorities of the expertise topic to obtain a second plurality of accounts as the authorities of the expertise topic, determining a third plurality of TGs in which a second number of accounts from the second plurality of accounts are members, determining that another frequent account is a member in one of the third plurality of TGs, and obtaining a ranking of accounts that are an authority on the expertise topic.

First claim

What is claimed is:

1. A method comprising:
identifying, for a first account, an interest of the account;
finding, in a topic ontology model, an expertise topic corresponding to the interest of the first account, wherein the topic ontology model is generated by:
obtaining data representing accounts that are associated with respective topics, each of the accounts being an account of a social network platform, each of the respective topics being an expertise topic of the account, and each of the accounts being associated with two or more respective topics; and
generating the topic ontology based on the expertise topics of the accounts, comprising:
ranking the topics associated with each account on a per account basis such that every account is associated with a ranked list of topics, wherein only a specified number of topics are retained for each account;
identifying all topics associated with a number of accounts that is less than a threshold number of accounts, and for each of the identified topics:
creating a child node for the identified topic;
determining a parent node for the child node, including determining a highest ranked topic other than the identified topic in each of the ranked lists in which the identified topic appears and creating a parent node for a most frequently occurring topic among the highest ranked topics;
generating a parent-child relationship between the parent node and the child node; and
removing the topic from the expertise topics of the accounts;

if there is more than one expertise topic per account, repeating the identifying until there is no more than one expertise topic per account; and
combining the parent-child relationships generated in the generating operations to generate the topic ontology model;

identifying one or more accounts associated with the expertise topic; and
providing one or more messages to the first account, wherein the one or more messages are from accounts that the first account does not have a graph relationship with, and wherein the one or more messages are from at least one of the identified one or more accounts associated with the expertise topic.

2. The method of claim 1 , wherein identifying one or more accounts associated with the expertise topic further comprises:
identifying a first plurality of accounts as top authorities for an expertise topic;
using the topic ontology model to identify one or more topic groups for the expertise topic in which a threshold number of accounts from the first plurality of accounts identified as authorities on the expertise topic are members;
selecting a frequent account that is also a member in at least one of the one or more topic groups in which the threshold number of accounts from the first plurality of accounts appear as members;
adding the frequent account as an authority on the expertise topic; and
iteratively adding additional frequent accounts as authorities on the expertise topic until a stopping criterion is reached, wherein for each iteration the frequent account of the prior iteration is added to the first plurality of accounts when identifying topic groups used to select a next frequent account.

3. The method of claim 1 , further comprising:
identifying, for the first account, a related topic from (i) interests of the account including a first interest of the first account and (ii) the topic ontology model, the related topic not being one of the interests of the first account, wherein identifying the related topic includes finding the related topic by finding in the topic ontology model a first node corresponding to the first interest of the account and a related node connected closely to the first node, wherein the related node is connected closely by virtue of being a direct parent of the first node, a direct grandparent of the first node, or a child of the direct parent or the direct grandparent;
generating a recommendation for the first account based on the related topic; and
providing the recommendation to the first account.

4. The method of claim 3 , wherein:
the recommendation is a recommendation to follow a second account that has the related topic as an expertise topic of the second account.

5. The method of claim 3 , wherein:
providing the recommendation comprises providing a message to the first account, wherein the message is from a third account that the first account does not have a graph relationship with, and wherein the third account has the related topic as an expertise topic of the third account.

6. The method of claim 1 , wherein the one or more accounts include at least one sponsored account and wherein providing one or more messages comprises providing an advertisement message from the at least one sponsored account, wherein the advertisement message corresponds to a product or service associated with the interest of the first account.

7. A system comprising:
a social network platform implemented on one or more computers and configured to connect to a plurality of client devices, each client device being associated with an account managed by the social network platform; and
an account repository comprising a plurality of accounts which are members of the social network platform,
the social network platform being configured to provide messages to the plurality of accounts by:
identifying, for a first account, an interest of the account;
finding, in a topic ontology model, an expertise topic corresponding to the interest of the first account, wherein the topic ontology model is generated by the social network platform by:
obtaining data representing accounts that are associated with respective topics, each of the accounts being an account of the social network platform, each of the respective topics being an expertise topic of the account, and each of the accounts being associated with two or more respective topics; and
generating the topic ontology based on the expertise topics of the accounts, comprising:
ranking the topics associated with each account on a per account basis such that every account is associated with a ranked list of topics, wherein only a specified number of topics are retained for each account;
identifying all topics associated with a number of accounts that is less than a threshold number of accounts, and for each of the identified topics:
creating a child node for the identified topic;
determining a parent node for the child node, including determining a highest ranked topic other than the identified topic in each of the ranked lists in which the identified topic appears and creating a parent node for a most frequently occurring topic among the highest ranked topics;
generating a parent-child relationship between the parent node and the child node; and
removing the topic from the expertise topics of the accounts;

if there is more than one expertise topic per account, repeating the identifying until there is no more than one expertise topic per account; and
combining the parent-child relationships generated in the generating operations to generate the topic ontology model;

identifying one or more accounts associated with the expertise topic; and
providing one or more messages to the first account, wherein the one or more messages are from accounts that the first account does not have a graph relationship with, and wherein the one or more messages are from at least one of the identified one or more accounts associated with the expertise topic.

8. The system of claim 7 , wherein identifying one or more accounts associated with the expertise topic further comprises:
identifying a first plurality of accounts as top authorities for an expertise topic;
using the topic ontology model to identify one or more topic groups for the expertise topic in which a th

Assignees

• Twitter Inc

Inventors

• Kolcz Alek

Classifications

• H04L67/306 Primary

  User profiles · CPC title

• H04L67/566

  Grouping or aggregating service requests, e.g. for unified processing · CPC title

• G06F16/335

  Filtering based on additional data, e.g. user or group profiles (filtering in web context G06F16/9535, G06F16/9536) · CPC title

• H04L67/04

  specially adapted for terminals or networks with limited capabilities; specially adapted for terminal portability · CPC title

• G06F16/24575 Primary

  using context · CPC title