Techniques for automatically sorting emails into folders

1. A method for automatically sorting emails into one or more folders managed by an email application, the method comprising:

receiving a new email;

establishing a first set of n-grams based on the new email;

establishing, for the new email, a set of high-value n-grams based on an intersection of (i) the first set of n-grams based on the new email, and (ii) the respective sets of high-value n-grams for the one or more folders;

establishing, for each folder, a respective affinity score based on (i) the respective set of high-value n-grams for the folder, and (ii) the set of high-value n-grams for the new email;

identifying a target folder among the one or more folders having a strongest affinity score; and

in response to identifying that the strongest affinity score satisfies a second threshold value:

associating the new email with the target folder.

2. The method of claim 1 , further comprising, prior to receiving the new email:

accessing a plurality of emails, wherein each email is included in a respective folder of the one or more folders;

establishing a second set of n-grams based on the plurality of emails;

for each n-gram in the second set of n-grams:

calculating a respective normalized entropy for the n-gram, and

calculating a respective indexing power for the n-gram based on its respective normalized entropy;

ranking the n-grams based on their respective indexing powers to establish a ranked list of n-grams; and

establishing, for each folder of the one or more folders, a respective set of high-value n-grams based on the ranked list of n-grams.

3. The method of claim 2 , wherein the respective normalized entropy for each n-gram is based on (i) a count of the one or more folders, and (ii) a sum, for each folder of the one or more folders, of: a first number of times the n-gram occurs in emails belonging to the folder relative to a second number of times the n-gram occurs within each email of the plurality of emails.

4. The method of claim 2 , wherein the respective indexing power for each n-gram is based on subtracting the respective normalized entropy for the n-gram from a value of one.

5. The method of claim 2 , further comprising, subsequent to establishing the ranked list of n-grams:

truncating the ranked list of n-grams in accordance with a first threshold value to cause at least one n-gram to be removed from the ranked list of n-grams.

6. The method of claim 5 , further comprising:

storing the at least one n-gram in a set of low-value n-grams.

7. The method of claim 2 , wherein each n-gram comprises one to N words derived from a subject, a body, or an attachment of at least one email of the plurality of emails.

8. The method of claim 1 , further comprising, in response to identifying that the strongest affinity score does not satisfy the second threshold value:

establishing a third set of n-grams based on a relative complement of (1) ( i ) the respective sets of high-value n-grams for the one or more folders, and (ii) the set of low-value n-grams, with respect to (2) the second first set of n-grams based on the new email, wherein the third set of n-grams includes one or more n-grams; and

in response to identifying, among the one or more n-grams, that at least one n-gram of the one or more n-grams is observed within the new email a number of times that satisfies a third threshold value:

issuing a prompt to create a new folder into which the new email can be placed, wherein the new folder is based on the at least one n-gram.

9. The method of claim 8 , further comprising:

in response to receiving an approval in association with the prompt:

creating the new folder, and

associating the new email with the new folder; or

in response to receiving a refusal in association with the prompt:

retaining the new email in a default folder into which new emails are placed.

10. At least one non-transitory computer readable storage medium configured to store instructions that, when executed by at least one processor included in a computing device, cause the computing device to automatically sort emails into one or more folders managed by an email application, by carrying out steps that include:

receiving a new email;

establishing a first set of n-grams based on the new email;

establishing, for the new email, a set of high-value n-grams based on an intersection of (i) the first set of n-grams based on the new email, and (ii) the respective sets of high-value n-grams for the one or more folders;

establishing, for each folder, a respective affinity score based on (i) the respective set of high-value n-grams for the folder, and (ii) the set of high-value n-grams for the new email;

identifying a target folder among the one or more folders having a strongest affinity score; and

in response to identifying that the strongest affinity score satisfies a second threshold value:

associating the new email with the target folder.

11. The at least one non-transitory computer readable storage medium of claim 10 , wherein the steps further include, prior to receiving the new email:

accessing a plurality of emails, wherein each email is included in a respective folder of the one or more folders;

establishing a second set of n-grams based on the plurality of emails;

for each n-gram in the second set of n-grams:

calculating a respective normalized entropy for the n-gram, and

calculating a respective indexing power for the n-gram based on its respective normalized entropy;

ranking the n-grams based on their respective indexing powers to establish a ranked list of n-grams; and

establishing, for each folder of the one or more folders, a respective set of high-value n-grams based on the ranked list of n-grams.

12. The at least one non-transitory computer readable storage medium of claim 11 , wherein the respective normalized entropy for each n-gram is based on (i) a count of the one or more folders, and (ii) a sum, for each folder of the one or more folders, of: a first number of times the n-gram occurs in emails belonging to the folder relative to a second number of times the n-gram occurs within each email of the plurality of emails.

13. The at least one non-transitory computer readable storage medium of claim 11 , wherein the respective indexing power for each n-gram is based on subtracting the respective normalized entropy for the n-gram from a value one.

14. The at least one non-transitory computer readable storage medium of claim 11 , wherein the steps further include, subsequent to

establishing the ranked list of n-grams:

truncating the ranked list of n grams in accordance with respect a first threshold value to cause at least one n-gram to be removed from the ranked list of n-grams.

15. The at least one non-transitory computer readable storage medium of claim 10 , wherein the steps further include, in response to identifying that the strongest affinity score does not satisfy the second threshold value:

establishing a third set of n-grams based on a relative complement of (1) ( i ) the respective sets of high-value n-grams for the one or more folders, and (ii) the set of low-value n-grams, with respect to (2) the first set of n-grams based on the new email, wherein the third set of n-grams includes one or more n-grams; and

in response to identifying, among the one or more n-grams, that at least one n-gram of the one or more n-grams is observed within the new email a default number of times that satisfies a third threshold value:

issuing a prompt to create a new folder into which the new email can be placed, wherein the new folder is based on the at least one n-gram.

16. A computing device configured to automatically sort emails into one or more folders managed by an email application, the computing device comprising:

at least one processor;

and at least one memory storing instructions that, when executed by the at least one processor, cause the computing device to:

receive a new email;

establish a first set of n-grams based on the new email;

establish, for the new email, a set of high-value n-grams based on an intersection of (i) the first set of n-grams based on the new email, and (ii) the respective sets of high-value n-grams for the one or more folders;

establish, for each folder a respective affinity score based on (i) the respective set of high-value n-grams for the folder, and (ii) the set of high-value n-grams for the new email;

identify a target folder among the one or more folders having a strongest affinity score; and

in response to identifying that the strongest affinity score satisfies a second threshold value:

associate the new email with the target folder.

17. The computing device of claim 16 , wherein the at least one processor further causes the computing device to, prior to receiving the new email:

accessing a plurality of emails, wherein each email is included in a respective folder of the one or more folders;

establishing a second set of n-grams based on the plurality of emails;

for each n-gram in the second set of n-grams:

calculating a respective normalized entropy for the n-gram, and

calculating a respective indexing power for the n-gram based on its respective normalized entropy;

ranking the n-grams based on their respective indexing powers to establish a ranked list of n-grams; and

establishing, for each folder of the one or more folders, a respective set of high-value n-grams based on the ranked list of n-grams.

18. The computing device of claim 17 , wherein the respective normalized entropy for each n-gram is based on (i) a count of the one or more folders, and (ii) a sum, for each folder of the one or more folders, of: a first number of times the n-gram occurs in emails belonging to the folder relative to a second number of times the n-gram occurs within each email of the plurality of emails.

19. The computing device of claim 17 , wherein the respective indexing power for each n-gram is based on subtracting the respective normalized entropy for the n-gram from a value of one.

20. The computing device of claim 17 , wherein the at least one processor further causes the computing device to, in response to identifying that the strongest affinity score does not satisfy the second threshold value:

establish a third set of n-grams based on a relative complement of (1) ( i ) the respective sets of high-value n-grams for the one or more folders, and (ii) the set of low-value n-grams, with respect to (2) the first set of n-grams based on the new email, wherein the third set of n-grams includes one or more n-grams; and

in response to identifying, among the one or more n-grams, that at least one n-gram of the one or more n-grams is observed within the new email a number of times that satisfies a third threshold value:

issue a prompt to create a new folder into which the new email can be placed, wherein the new folder is based on the at least one n-gram.