Adaptive notification

1. An apparatus comprising:

a processor;

a memory that stores code executable by the processor to:

determine a message sentiment of a message;

determine a sender relationship of a sender of the message to a recipient based on a message frequency MF calculated as MF=k(NM+(NR/avg(PT-RT)))/NM, wherein NM is a number of messages, NR is a number of reply messages, PT is a reply timestamp, RT is a receive timestamp, and k is a nonzero constant;

determine a frequency score FS calculated as FS=k√{square root over (MF)}, wherein MF is the message frequency;

determine a notification urgency for the message to the recipient based on a notification model based on the message sentiment the frequency score, and the sender relationship; and

communicate a notification of the message to the recipient based on the notification urgency.

2. The apparatus of claim 1 , wherein the notification model is a neural network recursively trained with the message sentiment, the sender relationship, and an urgency feedback.

3. The apparatus of claim 1 , wherein the notification model scores the message sentiment and the sender relationship and calculates the notification urgency as a function of the message sentiment and the sender relationship.

4. The apparatus of claim 1 , wherein the notification urgency specifies one or more of an alert volume, an alert tone, an alert tune, an alert duration, a vibration intensity, and a vibration duration.

5. The apparatus of claim 1 , wherein the message frequency is further calculated as a function of messages between the sender and the recipient.

6. The apparatus of claim 1 , wherein the code is further executable by the processor to determine a recipient situation, wherein the notification model is further based on the recipient situation.

7. The apparatus of claim 6 , wherein the recipient situation comprises one or more of a recipient location, a recipient action, a time frame, and a recipient status.

8. A method comprising:

determining, by use of a processor, a message sentiment of a message;

determining a sender relationship of a sender of the message to a recipient based on a message frequency MF calculated as MF=k(NM+(NR/avg(PT-RT)))/NM, wherein NM is a number of messages, NR is a number of reply messages, PT is a reply timestamp, RT is a receive timestamp, and k is a nonzero constant;

determining a frequency score FS calculated as FS=k√{square root over (MF)}, wherein MF is the message frequency;

determining a notification urgency for the message to the recipient based on a notification model based on the message sentiment, the frequency score, and the sender relationship; and

communicating a notification of the message to the recipient based on the notification urgency.

9. The method of claim 8 , wherein the notification model is a neural network recursively trained with the message sentiment, the sender relationship, and an urgency feedback.

10. The method of claim 8 , wherein the notification model scores the message sentiment and the sender relationship and calculates the notification urgency as a function of the message sentiment and the sender relationship.

11. The method of claim 8 , wherein the notification urgency specifies one or more of an alert volume, an alert tone, an alert tune, an alert duration, a vibration intensity, and a vibration duration.

12. The method of claim 8 , wherein the message frequency is further calculated as a function of messages between the sender and the recipient.

13. The method of claim 8 , the method further comprising determining a recipient situation, wherein the notification model is further based on the recipient situation.

14. The method of claim 13 , wherein the recipient situation comprises one or more of a recipient location, a recipient action, a time frame, and a recipient status.

15. A program product comprising a non-transitory computer readable storage medium that stores code executable by a processor, the executable code comprising code to:

determine a message sentiment of a message;

determine a sender relationship of a sender of the message to a recipient based on a message frequency MF calculated as MF=k(NM+(NR/avg)PT-RT)))/NM, wherein NM is a number of messages, NR is a number of reply messages, PT is a reply timestamp, RT is a receive timestamp, and k is a nonzero constant;

determine a frequency score FS calculated as FS=k√{square root over (MF)}, wherein MF is the message frequency;

determine a notification urgency for the message to the recipient based on a notification model based on the message sentiment, the frequency score, and the sender relationship; and

communicate a notification of the message to the recipient based on the notification urgency.

16. The program product of claim 15 , wherein the notification model is a neural network recursively trained with the message sentiment, the sender relationship, and an urgency feedback.

17. The program product of claim 15 , wherein the notification model scores the message sentiment and the sender relationship and calculates the notification urgency as a function of the message sentiment and the sender relationship.

18. The program product of claim 15 , wherein the notification urgency specifies one or more of an alert volume, an alert tone, an alert tune, an alert duration, a vibration intensity, and a vibration duration.

19. The program product of claim 15 , wherein the message frequency is further calculated as a function of messages between the sender and the recipient.

20. The program product of claim 15 , wherein the code is further executable by the processor to determine a recipient situation, wherein the notification model is further based on the recipient situation.