Browse Prior Art Database

Identifying Entities That Hang Out With Each Other

IP.com Disclosure Number: IPCOM000239413D
Publication Date: 2014-Nov-05
Document File: 5 page(s) / 44K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a hangout-detecting software application that monitors the movement and/or communication/interaction behavior of entities, flagging conditions where an entity is observed to hang out with another entity.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 23% of the total text.

Page 01 of 5

Identifying Entities That Hang Out With Each Other

This article relates to the field of software-implemented behavioral analytics, including event-driven observation and tracking of entities for the purpose of flagging occurrences of certain repeating behavior of the tracked entities.

Entity analytics software can identify entities as related by common features (e.g., proximity in space and time). However, some entity interactions become of interest only upon repeating. Two entities that occasionally happen to pass by one another may have no significant relationship. Entities that repeatedly meet in various locations within a short timeframe are likely to have a reason for those meetings. Meetings of entities, either physically or via a communication medium, may indicate a kind of entity relationship that warrants identification. Meetings that repeat, and/or meetings that involve a significant number of entities, may be worthwhile to distinguish from other meetings of entities.

Entities that physically hang out with each other can be identified when a first entity is observed within a certain spatial proximity of a second entity, on at least a certain number of occasions, over a certain time interval. Entities that virtually hang out with each other can be identified when a first entity is observed to communicate or otherwise interact with a second entity, on at least a certain number of occasions, over a certain time interval. Entities that hang out on an ongoing basis, or "friends", can be identified based on a required number of observations of hangouts. Entities that hang out with each other in large groups, or "gatherings", can be identified based simply on a required number of involved entities; a gathering can be identified either when a certain number of entities is observed within a certain physical distance of each other, or when a certain number of entities virtually interact with each other.

The novel contribution is detection of these interactions regardless of ongoing changes in spatial location. In one embodiment, the application monitors the movement and/or communication/interaction behavior of entities, flagging conditions in which an entity is observed hanging out with another entity. Multiple entities can be flagged to hang out together. The data sources, entity types, geographic areas, and communication/ interaction techniques covered by hangout detection can be user-configurable. The hangout detector can run as an operator for analytics platforms or as a standalone executable. It can use in-memory event tracking to detect hangouts with optimum efficiency.

In-memory hangout detector design considerations

For performance at scale, the hangout detector can be implemented using multiple computing nodes that receive data about observed events and maintain that data in memory. Because the hangout detector's in-memory event data is not shared across processes, incoming behavioral data for a particular entity must be...