Systems and methods for node resolution using multiple fields with dynamically determined priorities based on field values

Patent metadata

Field	Value
Publication number	US-10516587-B2
Application number	US-201916360884-A
Country	US
Kind code	B2
Filing date	Mar 21, 2019
Priority date	May 24, 2018
Publication date	Dec 24, 2019
Grant date	Dec 24, 2019

Abstract

The systems and methods described herein can use multiple fields with dynamically determined priorities based on field values for node resolution. The system can generate activity field-value pairs including an activity value associated with an activity field from an electronic activity. The system can determine a frequency score based on a first count of node field-value pairs that match the activity value. The system can assign a weight to the activity value based on the frequency score. The system can generate a match score of a candidate node profile indicating a likelihood that the electronic activity is transmitted or received by an account corresponding to the candidate node profile. The system can store an association between the electronic activity and the node profile selected based on the match score.

First claim

What is claimed is:

1. A method, comprising:
maintaining, by one or more processors, a plurality of node profiles, each node profile of the plurality of node profiles corresponding to a respective entity and including node field-value pairs;
accessing, by the one or more processors, data of a plurality of electronic activities transmitted or received via at least one electronic account associated with at least one data source provider;
generating, by the one or more processors, a plurality of activity field-value pairs from data of an electronic activity of the plurality of electronic activities, each activity field-value pair of the plurality of activity field-value pairs including a respective activity value associated with a respective activity field;
determining, by the one or more processors, for a first node field-value pair that matches a first activity field-value pair of the plurality of activity field-value pairs, a first count corresponding to a first number of node profiles of the plurality of node profiles that includes the first node field-value pair;
determining, by the one or more processors, for a second node field-value pair that matches a second activity field-value pair of the plurality of activity field-value pairs, a second count corresponding to a second number of node profiles of the plurality of node profiles that includes the second node field-value pair;
determining, by the one or more processors, a first frequency score of the first activity field-value pair based on the first count and a second frequency score of the second activity field-value pair based on the second count;
assigning, by the one or more processors, a first weight to the first activity field-value pair based on the first frequency score and a second weight to the second activity field-value pair based on the second frequency score, the first weight greater than the second weight responsive to the first frequency score being less than the second frequency score;
generating, by the one or more processors, for each candidate node profile of a plurality of candidate node profiles of the plurality of node profiles, a respective match score of the candidate node profile indicating a likelihood that the electronic activity is transmitted or received by an electronic account corresponding to the candidate node profile based on the first weight assigned to the first activity field-value pair and the second weight assigned to the second activity field-value pair;
selecting, by the one or more processors, from the plurality of candidate node profiles, a node profile based on the respective match score of the node profile; and
storing, by the one or more processors, in one or more data structures, an association between the electronic activity and the node profile.

2. The method of claim 1 , wherein the electronic activity is a first electronic activity, the method further comprising:
accessing, by the one or more processors, data of a second electronic activity;
determining, by the one or more processors, for a third node field-value pair that matches a third activity field-value pair generated from the second electronic activity, a third count corresponding to a third number of node profiles of the plurality of node profiles that includes the third node field-value pair;
determining, by the one or more processors, for a fourth node field-value pair that matches a fourth activity field-value pair generated from the second electronic activity, a fourth count corresponding to a fourth number of node profiles of the plurality of node profiles that includes the fourth node field-value pair;
determining, by the one or more processors, a third frequency score of the third activity field-value pair of the second electronic activity based on the third count and a fourth frequency score of the fourth activity field-value pair of the second electronic activity based on the fourth count, the third activity field-value pair associated with the first activity field-value pair, the fourth activity field-value pair associated with the second activity field-value pair; and
assigning, by the one or more processors, a third weight to the third activity field-value and a fourth weight to the fourth activity field-value pair, the third weight less than the fourth weight responsive to the third frequency score being greater than the fourth frequency score.

3. The method of claim 1 , further comprising:
determining, by the one or more processors, at least one of a first location or a first time zone of a participant of the electronic activity; and
wherein generating, for each candidate node profile, the respective match score comprises generating, by the one or more processors, the match score of the node profile based on comparing the at least one of the first location or the first time zone to at least one of a second location or a second time zone associated with the node profile.

4. The method of claim 1 , further comprising:
maintaining, by the one or more processors, at least one of a first name distribution based on a count of each node value of the plurality of node field-value pairs that is associated with a first name field or a last name distribution based on a count of each node value of the plurality of node field-value pairs that is associated with a last name field.

5. The method of claim 4 , further comprising:
determining, by the one or more processors, the first frequency score or the second frequency score based on the at least one of the first name distribution or the last name distribution.

6. The method of claim 4 , further comprising:
generating, by the one or more processors, the plurality of activity field-value pairs using at least one of the first name distribution or the last name distribution.

7. The method of claim 1 , further comprising:
determining, by the one or more processors, at least one of (i) a location associated with a particular node profile of the plurality of node profiles based on location data associated with second electronic activities linked to the particular node profile or (ii) a time zone of the particular node profile based on timestamp data associated with the second electronic activities linked to the particular node profile; and
wherein generating, for each candidate node profile, the respective match score comprises generating, by the one or more processors, the match score for the particular node profile based on the at least one of the location or the time zone of the particular node profile.

8. A system comprising:
one or more hardware processors configured by machine-readable instructions to:
maintain a plurality of node profiles, each node profile of the plurality of node profiles corresponding to a respective entity and including node field-value pairs;
access data of a plurality of electronic activities transmitted or received via at least one electronic account associated with at least one data source provider;
generate a plurality of activity field-value pairs from data of an electronic activity of the plurality of electronic activities, each activity field-value pair of the plurality of activity field-value pairs including a respective activity value associated with a respective activity field;
determine, for a first node field-value pair that matches a first activity field-value pair of the plurality of activity field-value pairs, a first count corresponding to a first number of node profiles of the plurality of node profiles that includes the first node field-value pair;
determine, for a second node field-value pair that matches a second activity field-value pair of the plurality of activity field-value pairs, a second count corresponding to a second number of node profiles of the plurality of node p

Assignees

• People Ai Inc

Inventors

• Brunets Yurii
• Rice Devin
• Rogynskyy Oleg
• Wilson James

Classifications

• G06N5/025

  Extracting rules from data · CPC title

• G06N5/022

  Knowledge engineering; Knowledge acquisition · CPC title

• G06N20/00

  Machine learning · CPC title

• H04L67/30

  Profiles · CPC title

• G06Q10/10

  Office automation; Time management · CPC title