Method and system for activity based email sorting

• US 7,587,461 B2
• Filed: 04/19/2005
• Issued: 09/08/2009
• Est. Priority Date: 04/19/2005
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Rowley, Henry A., Baluja, Shumeet
• Primary Examiner(s): Follansbee; John
• Assistant Examiner(s): WIDHALM, ANGELA M
• Application Number:
  US11/110,116
• Publication Number:
  US 20060235933A1
• Time in Patent Office: 1,603 Days
• Field of Search: None
• US Class Current: 709/207
• CPC Class Codes: G06Q 10/107 Computer-aided management of electronic mailing [e-mailing]
• CPC Class Codes: H04L 51/226 Delivery according to priorities

Abstract

Messages of a user are ranked based on metrics derived from user actions with regard to the messages. Groupings of messages, such as conversations or a set of messages that are determined to be similar, may also be ranked. When new messages are received, their interest to the user is estimated based on rankings of messages and message groupings to which the new message is related. The new messages may be made more or less prominent based on the estimated interest to the user.

Claims

1. A method comprising:

on a message server computer having one or more processors and memory,

for each conversation i in a plurality of n conversations (i=1, . . . , n), wherein each conversation i includes messages, wherein the messages include messages received by a user;

determining a respective received message ranking value v_j for each received message in the conversation, wherein;

the respective received message ranking value v_j is a function of at least one interaction metric m_k for the respective received message;

v_j=ƒ

(m_k), wherein

m_k is a value of the interaction metric for a k-th interaction metric, and

the interaction metric m_k is based on a predefined action performed by the user with regard to the respective received message; and

determining a conversation ranking value r_i for the conversation, wherein the conversation ranking value r_i is a function of received message ranking values v_j for all of the received messages in the conversation;

r_i=ƒ

(v_j), (j=1, . . . , m), wherein

r_i is the conversation ranking value for an i-th conversation in the plurality of n conversations, and

v_j is the message ranking value for a j-th received message in the plurality of m received messages in the i-th conversation; and

sorting the conversations in the plurality of n conversations in accordance with the determined conversation ranking values r_i, (i=1, . . . , n).

2. The method of claim 1, wherein the messages in a respective conversation include email messages.

3. The method of claim 1, wherein determining at least one interaction metric for a respective received message comprises determining a length of time between a time of receipt of the received message and a time that the user performs a predefined action with respect to the received message.

4. The method of claim 3, wherein the predefined action includes one of:

deleting the received message, responding to the received message, forwarding the received message, and reading the received message.

5. The method of claim 1, wherein said determining at least one interaction metric for a respective received message comprises determining a number of times the user has performed a predefined action with respect to the received message, wherein the predefined action includes one of:

responding to the received message, forwarding the received message, and reading the received message.

6. The method of claim 1, wherein each received message is assigned a respective received message ranking value v_j based on a weighted average of a plurality of p interaction metrics for the received message:

$v_{j<br><br>}=\underset{k=1}{\overset{p}{\sum <br><br>}}<br><br><br><br>\left(\left({\mathrm{\alpha <br><br>}}_k<br><br>m_k\right)/p\right),$

wherein

m_k is a value of the interaction metric for a k-th interaction metric in the plurality of p interaction metrics, and

α

_k are respective weights for respective interaction metrics.

7. The method of claim 1, including assigning the respective received message ranking value to a message in accordance with a linear combination of the interaction metrics for the message:

$v_j=\underset{k=1}{\overset{p}{\sum <br><br>}}<br><br><br><br>{\mathrm{\alpha <br><br>}}_k<br><br>m_k,$

wherein

m_k is a value of the interaction metric for a k-th interaction metric in the plurality of p interaction metrics, and

α

_k are respective weights for respective interaction metrics.

8. The method of claim 1, wherein sorting the conversations in the plurality of conversations includes ordering the conversations in the plurality of conversations in accordance with the determined conversation ranking values.

16. The method of claim 1, wherein each received message ranking value v_j is a function of a plurality of p interaction metrics for the corresponding received message:

v_j=ƒ

(m_k) wherein

v_j is the message ranking value for a j-th received message in the plurality of m received messages in the i-th conversation, and

m_k is a value of the interaction metric for a k-th interaction metric in the plurality of p interaction metrics.

17. The method of claim 1, wherein the received message ranking values v_j are weighted based on receipt times of the received messages in the conversation.

18. The method of claim 1, wherein the conversation ranking value r_i is a linear combination of the received message ranking values v_j for all of the received messages in the conversation:

$r_i=\underset{j=1}{\overset{m}{\sum <br><br>}}<br><br><br><br>{\mathrm{\beta <br><br>}}_j<br><br>v_j$

wherein

r_i is the conversation ranking value for an i-th conversation in the plurality of n conversations,

v_j is the message ranking value for a j-th received message in the plurality of m received messages in the i-th conversation, and

β

_j is the respective weight for a j-th message in the plurality of m received messages in the i-th conversation.

19. The method of claim 1, wherein the conversation ranking value is a weighted average of the received message ranking values for all of the received messages in the conversation:

$r_i=\underset{j=1}{\overset{m}{\sum <br><br>}}<br><br><br><br>\left(\left({\mathrm{\beta <br><br>}}_j<br><br>v_j\right)/m\right)$

wherein

r_i is the conversation ranking value for an i-th conversation in the plurality of n conversations,

vj is the message ranking value for a j-th received message in the plurality of m received messages in the i-th conversation, and

β

_j is the respective weight for a j-th message in the plurality of m received messages in the i-th conversation.

20. The method of claim 1, wherein the conversation ranking value is an arithmetic mean of the received message ranking values v_j for all of the received messages in the conversation:

$r_i=\underset{j=1}{\overset{m}{\sum <br><br>}}<br><br><br><br>\left(v_j\right)/m$

wherein

r_i is the conversation ranking value for an i-th conversation in the plurality of n conversations, and

v_j is the message ranking value for a j-th received message in the plurality of m messages in the i-th conversation.

21. The method of claim 7, wherein a respective weight α

_k for a respective interaction metric decreases with time based on how long ago a respective message was received.

22. The method of claim 21, wherein α

_k ∝

1/t, where t is the time since the respective message was received.

23. The method of claim 18, wherein a respective weight β

_j for a respective received message decreases with time based on how long ago the respective message was received.

24. The method of claim 23, wherein β

_j ∝

1/t, where t is the time since the respective message was received.

25. The method of claim 23, wherein the respective weight β

_j exponentially decays over time based on how long ago the respective message was received.

9. A computer system for ordering a plurality of conversations, the system, comprising:

one or more processors;

memory;

a new message processing module to order a plurality of n conversations, each conversation i (i=1, . . . , n) including messages, wherein the messages include messages received by a user; and

a message ranking module to determine a respective received message ranking value v_j for each received message in the conversation, wherein;

the respective received message ranking value v_j is a function of at least one interaction metric m_k for the respective received message;

v_j=ƒ

(m_k),

wherein

m_k is a value of the interaction metric for a k-th interaction metric and

the interaction metric m_k is based on at least one predefined action performed by the user with regard to the respective received message; and

a conversation ranking module to determine a conversation ranking value r_i of the conversation, wherein the conversation ranking value r_i is a function of received message ranking values v_j for all of the received messages in the conversation;

r_i=ƒ

(v_j), (j=1, . . . , m),

wherein

r_i is the conversation ranking value for an i-th conversation in the plurality of n conversations, and

v_j is the message ranking value for a j-th received message in the plurality of m received messages in the i-th conversation.

10. The computer system of claim 9, wherein the message ranking module includes instructions to determine a length of time between a time of receipt of a respective received message and a time that the user performs a predefined action with respect to the received message, wherein the predefined action includes one of:

deleting the received message, forwarding the received message, reading the received message, and responding to the received message.

11. The computer system of claim 9, wherein the message ranking module includes instructions to determine a number of times the user has performed a predefined action with respect to a respective received message, wherein the predefined action includes one of:

responding to the received message, forwarding the received message, and reading the received message.

12. The computer system of claim 9, wherein the message ranking module includes instructions to assign a message ranking value v_j to each received message based on a weighted average of a plurality of p interaction metrics for the received message:

$v_{j<br><br>}=\underset{k=1}{\overset{p}{\sum <br><br>}}<br><br><br><br>\left(\left({\mathrm{\alpha <br><br>}}_k<br><br>m_k\right)/p\right),$

wherein

m_k is a value of the interaction metric for a k-th interaction metric in the plurality of p interaction metrics, and

α

_k are respective weights for respective interaction metrics.

13. The computer system of claim 9, wherein the message ranking module includes instructions to assign a message ranking value to a respective message in accordance with a linear combination of the interaction metrics of the message:

$v_j=\underset{k=1}{\overset{p}{\sum <br><br>}}<br><br><br><br>{\mathrm{\alpha <br><br>}}_k<br><br>m_k,$

wherein

m_k is a value of the interaction metric for a k-th interaction metric in the plurality of p interaction metrics, and

α

_k are respective weights for respective interaction metrics.

14. A computer readable storage medium and one or more computer programs embedded therein, the one or more computer programs comprising instructions, which when executed by a computer system, cause the computer system to:

for each conversation i in a plurality of n conversations (i=1, . . . , n), wherein each conversation i includes messages, wherein the messages include messages received by a user;

determine a respective received message ranking value v_j for each received message in the conversation, wherein;

the respective received message ranking value v_j is a function of at least one interaction metric m_k for the respective received message;

v_j=ƒ

(m_k), wherein

m_k is a value of the interaction metric for a k-th interaction metric, and

the interaction metric m_k is based on a predefined action performed by the user with regard to the respective received message; and

determine a conversation ranking value r_i for the conversation, wherein the conversation ranking value r_i is a function of received message ranking values v_j for all of the received messages in the conversation;

r_i=ƒ

(v_j), (j=1, . . . , m), wherein

r_i is the conversation ranking value for an i-th conversation in the plurality of n conversations, and

v_j is the message ranking value for a j-th received message in the plurality of m received messages in the i-th conversation; and

sort the conversations in the plurality of n conversations in accordance with the determined conversation ranking values r_i, (i=1, . . . , n).

15. A computer system, comprising:

one or more processors;

memory; and

one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including;

for each conversation i in a plurality of n conversations (i=1, . . . , n), wherein each conversation i includes messages, wherein the messages include messages received by a user;

instructions for determining a respective received message ranking value v_jfor each received message in the conversation, wherein;

the respective received message ranking value v_j is a function of at least one interaction metric m_k for the respective received message;

v_j=ƒ

(m_k), wherein

m_k is a value of the interaction metric for a k-th interaction metric, and

the interaction metric m_k is based on a predefined action performed by the user with regard to the respective received message; and

instructions for determining a conversation ranking value r_i for the conversation, wherein the conversation ranking value r_i is a function of received message ranking values v_j for all of the received messages in the conversation;

r_i=ƒ

(v_j), (j=1, . . . , m), wherein

r_i is the conversation ranking value for an i-th conversation in the plurality of n conversations, and

v_j is the message ranking value for a j-th received message in the plurality of m received messages in the i-th conversation; and

instructions for sorting the conversations in the plurality of n conversations in accordance with the determined conversation ranking values r_i, (i=1, . . . , n).