Mining global email folders for identifying auto-folder tags
Embodiments are directed towards identifying auto-folder tags for messages by using a combinational optimization approach of bi-clustering folder names and features of messages based on relationship strengths. The combinational optimization approach of bi-clustering, generally, groups a plurality of folder names and a plurality of features into one or more metafolders to optimize a cost. The cost is based on an aggregate of cut relationship strengths, where a cut results when a relationship folder name and feature are grouped in separate metafolders. Furthermore, the plurality of folder names and the plurality of features are obtained by monitoring actions of a plurality of users, where the folder names are user generated folder names and features are from a plurality of messages. The metafolders may be used to tag new user messages with an auto-folder tag.
1. A method operating on a network device, the method comprising:
monitoring actions over a first set of messages by a plurality of users;
employing the monitored actions by the plurality of users to generate metafolders for use by a user in the plurality of users using a combinational optimization approach of bi-clustering both folder names and features of the messages based on each relationship strength between at least the folder names and the features, cutting relationships between a plurality of the folder names and a plurality of message features when a folder name of a relationship and a message feature of the relationship are in separate metafolders, and regenerating the metafolders until a cost of cutting is within a confidence level that is based on an aggregate of relationship strengths of the cut relationships; and
using the generated metafolders to automatically tag at least one message for the user with an auto-folder tag.
2. The method of claim 1 , wherein the monitoring actions further comprises obtaining a plurality of folder names and a plurality of features from the plurality of users.
3. The method of claim 1 , wherein the relationship strengths are based on a comparison of a feature frequency for a folder name with a feature frequency for a global system.
4. The method of claim 1 , wherein the relationship strengths are based on a log likelihood ratio that estimates a relationship strength based on a previously determined relationship strength.
5. The method of claim 1 , wherein the metafolders further comprise a hierarchy of metafolders based on a frequency of messages in the one or more metafolders.
6. A network device, comprising:
a transceiver to send and receive data over a network; and
a processor that is operative on the received data to perform actions, including:
monitoring the actions over a first set of messages by a plurality of users;
employing the monitored actions by the plurality of users to generate metafolders for use by a user in the plurality of users using a combinational optimization approach of bi-clustering both folder names and features of the messages based on each relationship strength between at least the folder names and the features, cutting relationships between a plurality of the folder names and a plurality of message features when a folder name of a relationship and a message feature of the relationship are in separate metafolders, and regenerating the metafolders until a cost of cutting is within a confidence level that is based on an aggregate of relationship strengths of the cut relationships; and
using the generated metafolders to automatically tag at least one message for the user with an auto-folder tag.
7. The network device of claim 6 , wherein the monitoring actions further comprises obtaining a plurality of folder names and a plurality of features from the plurality of users.
8. The network device of claim 6 , wherein the relationship strengths are based on a comparison of a feature frequency for a folder name with a feature frequency for a global system.
9. The network device of claim 6 , wherein the relationship strengths are based on a log likelihood ratio that estimates a relationship strength based on a previously determined relationship strength.
10. The network device of claim 6 , wherein the metafolders further comprise a hierarchy of metafolders based on a frequency of messages in the one or more metafolders.
11. A system, comprising:
at least one network device that manages a-communications over a network; and
one or more other network devices that are configured to perform actions, including:
monitoring the actions over a first set of messages by a plurality of users;
employing the monitored actions by the plurality of users to generate metafolders for use by a user in the plurality of users using a combinational optimization approach of bi-clustering both folder names and features of the messages based on each relationship strength between at least the folder names and the features, cutting relationships between a plurality of the folder names and a plurality of message features when a folder name of a relationship and a message feature of the relationship are in separate metafolders, and regenerating the metafolders until a cost of cutting is within a confidence level that is based on an aggregate of relationship strengths of the cut relationships; and
using the generated metafolders to automatically tag at least one message for the user with an auto-folder tag.
12. The system of claim 11 , wherein the monitoring actions further comprises obtaining the plurality of folder names and the plurality of features from the plurality of users.
13. The system of claim 11 , wherein the relationship strengths are based on a comparison of a feature frequency for a folder name with a feature frequency for a global system.
14. The system of claim 11 , wherein the relationship strengths are based on a log likelihood ratio that estimates a relationship strength based on a previously determined relationship strength.