Identification of similar users

Patent metadata

Field	Value
Publication number	US-9756478-B2
Application number	US-201514978841-A
Country	US
Kind code	B2
Filing date	Dec 22, 2015
Priority date	Dec 22, 2015
Publication date	Sep 5, 2017
Grant date	Sep 5, 2017

Abstract

A peer group of a user may be identified on the basis of geographic information and building occupancy information for the user and for possible peers of the user. Energy usage data for each peer group member may be compared with that of the user. A comparison result may be computed between each peer group member and the user. A representation of comparison results may be provided to the user.

First claim

The invention claimed is:

1. A computer-implemented method comprising:
identifying a peer group of a user on the basis of geographic information and building occupancy information for the user and for possible peers of the user;
comparing energy usage data for each peer group member with energy usage data for the user and computing a comparison result between each peer group member and the user; and
providing a representation of the comparison results to the user,
wherein the geographic information and building occupancy information is formatted as one or more feature vectors,
wherein the identification of the peer group comprises comparing of feature vectors of possible peers with feature vectors for the user, and
wherein the comparison of the feature vectors comprises:
computing a geographic distance d g between the user and the possible peer;
computing an occupancy distance d o between the user and the possible peer; and
computing an overall distance d between the user and the possible peer, where d is a function of d g and d o .

2. The method of claim 1 , wherein the geographic information comprises:
location information; and
climate information.

3. The method of claim 1 , wherein the building occupancy information comprises a likelihood that a person is present in a building at each of a plurality of time intervals during a day.

4. The method of claim 1 , wherein the geographic distance d g is computed
d g ( x,y )=haversine( x 1 ,y 1 )+α∥ x c −y c ∥ for the user x and a possible peer y,

where x 1 and y 1 represent locations of the user x and the possible peer y respectively, x c and y c are feature vectors representing climate information for the user and the peer group member respectively, and α is a climate information weighting factor.

5. The method of claim 1 , wherein the occupancy distance d o is computed
d o ( x,y )=< x o ,y 0 >/∥x o ∥ 2 ∥y o ∥ 2 for the user x and a possible peer y,

where x o and y o are feature vectors representing the likelihood of occupancy in each of several time intervals by the user x in a building associated with user x, and by a possible peer y in a building associated with possible peer y, respectively, ∥ is a norm calculation, and < > is a dot product calculation.

6. The method of claim 1 , wherein the computing of the overall distance d comprises computing d(x, y)=d g +βd o , where β is an occupancy distance weighting factor.

7. The method of claim 6 , wherein the peer group comprises possible peers whose respective overall distances from the user are less than a predefined threshold distance.

8. The method of claim 6 , wherein the peer group comprises a predefined number of possible peers having the shortest overall distances from the user, of the possible peers.

9. A system, comprising:
a programmable processor; and
a memory in communication with the processor, the memory storing instructions which, when executed by the processor, cause the processor to:
identify a peer group of a user on the basis of geographic information and building occupancy information of the user and of possible peers of the user;
compare energy usage data for each peer group member with energy usage data for the user and computing a comparison result between each peer group member and the user; and

provide a representation of the comparison results to the user;
wherein the geographic information and building occupancy information is formatted as one or more feature vectors, and

wherein the instructions which cause the processor to identify the peer group cause the processor to compare feature vectors of possible peers with feature vectors for the user, wherein the comparison of the feature vectors comprises:
computing a geographic distance d g between the user and the possible peer;
computing an occupancy distance d o between the user and the possible peer; and
computing an overall distance d between the user and the possible peer, where d is a function of d g and d o .

10. The system of claim 9 , wherein the geographic information comprises:
location information; and
climate information.

11. The system of claim 9 , wherein the building occupancy information comprises likelihood that a person is present in a building at each of a plurality of time intervals during a day.

12. The system of claim 9 , wherein the geographic distance d g is computed
d g ( x,y )=haversine( x 1 ,y 1 )+α∥ x c −y c ∥ for the user x and a possible peer y,

where x 1 and y 1 represent locations of the user x and a possible peer y respectively, x c and y c are feature vectors representing climate information for the user and the possible peer respectively, and α is a climate information weighting factor.

13. The system of claim 9 , wherein the occupancy distance d o is computed
d o ( x,y )=< x o ,y o >/∥x o ∥ 2 ∥y o ∥ 2 for the user x and a possible peer y,

where x o and y o are feature vectors representing the likelihood of occupancy in each of several time intervals by the user x in a building associated with user x, and by a possible peer y in a building associated with peer group member y, respectively, ∥ is a norm calculation, and < > is a dot product calculation.

14. The system of claim 9 , wherein the computing of the overall distance d comprises computing d(x, y)=d g +βd o , where β is an occupancy distance weighting factor.

15. The system of claim 14 , wherein the peer group comprises possible peers whose respective overall distances from the user are less than a predefined threshold distance.

16. The system of claim 14 , wherein the peer group comprises a predefined number of possible peers having the shortest overall distances from the user, of the possible peers.

Assignees

• Google Inc

Inventors

• Shafi Sayed Yusef
• Hester Todd
• Ben-Meshulam Jonathan
• Dey Sourav Raj

Classifications

• H04W4/08 Primary

  User group management · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N20/00

  Machine learning · CPC title

• H04W4/029

  Location-based management or tracking services · CPC title

• H04L43/045

  for graphical visualisation of monitoring data · CPC title