Method and system for automatic formality classification

1. A method, implemented on a machine having at least one processor, storage, and a communication platform connected to a network for recommending a text item to an online user, the method comprising:

machine-training a plurality of models based on parallel textual data including different versions of training sentences each having a same literal meaning but different formality levels;

receiving a request for recommending one or more text items of a plurality of text items to an online user;

extracting one or more linguistic features from each of the plurality of text items;

extracting contextual information with respect to the online user;

determining, using a first model of the plurality of models, a formality level of each of the plurality of text items based on the one or more linguistic features and the contextual information, wherein the formality level represents a degree of formality of the text item;

determining an identity of the online user based on the contextual information;

determining the formality preference of the online user based on a user profile associated with the identity;

ranking the plurality of text items based on their respective formality levels and the formality preference of the online user indicative of a degree of formality preferred by the online user;

selecting the one or more text items from the plurality of text items based on the ranking;

recommending the one or more text items to the online user; and

updating at least one of the plurality of models based on the one or more linguistic features and the contextual information.

2. The method of claim 1 , wherein the formality preference of the online user is determined based on historical online behaviors of the online user.

3. The method of claim 1 , further comprising:

determining a granularity level with respect to each of the plurality of text items based on the formality preference of the online user, wherein the formality level of each of the plurality of text items is determined in accordance with the granularity level.

4. The method of claim 3 , wherein:

the granularity level of each of the plurality of text items is further determined based on a purpose of an online communication with which the one or more text items are to be included;

a first granularity level is used for determining the formality level in response to the purpose of the online communication being a first purpose; and

a second granularity level is used for determining the formality level in response to the purpose of the online communication being a second purpose.

5. The method of claim 1 , wherein the formality level is determined based on the first model trained with linguistic features of the parallel textual data in online communications and purposes associated with the online communications.

6. The method of claim 1 , wherein the formality level is further determined based on (i) a purpose of an online communication the one or more text items are to be included within, and (ii) a granularity level of each of the plurality of text items, wherein the granularity level represents a scale from at least one word to at least one paragraph at which the formality level of the text item is to be determined.

7. The method of claim 6 , wherein the purpose of the online communication is determined based on the contextual information.

8. The method of claim 1 , further comprising:

modifying the user profile associated with the online user based on the formality level.

9. The method of claim 1 , further comprising:

receiving training data comprising a plurality of online communications;

extracting linguistic features of textual data in the plurality of online communications from the training data;

extracting contextual information from the training data, wherein the contextual information from the training data comprises purposes associated with the plurality of online communications, and the purposes comprise at least a first purpose and a second purpose; and

employing the training data, the linguistic features extracted from the training data, and the purposes indicated by the contextual information extracted from the training data to generate or update a formality classification model used to determine the formality level of the plurality of text items.

10. A non-transitory computer-readable medium storing computer program instructions that, when executed by one or more processors, effectuate operations comprising:

machine-training a plurality of models based on parallel textual data including different versions of training sentences each having a same literal meaning but different formality levels;

receiving a request for recommending one or more text items of a plurality of text items to an online user;

extracting one or more linguistic features from each of the plurality of text items;

extracting contextual information with respect to the online user;

determining, using a first model of the plurality of models, a formality level of each of the plurality of text items based on the one or more linguistic features and the contextual information, wherein the formality level represents a degree of formality of the text item;

determining an identity of the online user based on the contextual information;

determining the formality preference of the online user based on a user profile associated with the identity;

ranking the plurality of text items based on their respective formality levels and the formality preference of the online user indicative of a degree of formality preferred by the online user;

selecting the one or more text items from the plurality of text items based on the ranking;

recommending the one or more text items to the online user; and

updating at least one of the plurality of models based on the one or more linguistic features and the contextual information.

11. The non-transitory computer-readable medium of claim 10 , wherein the formality preference of the online user is determined based on historical online behaviors of the online user.

12. The non-transitory computer-readable medium of claim 10 , wherein the operations further comprise:

determining a granularity level with respect to each of the plurality of text items based on the formality preference of the online user, wherein the formality level of each of the plurality of text items is determined in accordance with the granularity level.

13. The non-transitory computer-readable medium of claim 12 , wherein:

the granularity level of each of the plurality of text items is further determined based on a purpose of an online communication with which the one or more text items are to be included;

a first granularity level is used for determining the formality level in response to the purpose of the online communication being a first purpose; and

a second granularity level is used for determining the formality level in response to the purpose of the online communication being a second purpose.

14. The non-transitory computer-readable medium of claim 10 , wherein the formality level is determined based on the first model trained with linguistic features of the parallel textual data in online communications and purposes associated with the online communications.

15. The non-transitory computer-readable medium of claim 10 , wherein:

the formality level is further determined based on (i) a purpose of an online communication the one or more text items are to be included within, and (ii) a granularity level of each of the plurality of text items;

the granularity level represents a scale from at least one word to at least one paragraph at which the formality level of the text item is to be determined; and

the purpose of the online communication is determined based on the contextual information.

16. The non-transitory computer-readable medium of claim 10 , wherein the operations further comprise:

modifying the user profile associated with the online user based on the formality level.

17. The non-transitory computer-readable medium of claim 10 , wherein the operations further comprise:

receiving training data comprising a plurality of online communications;

extracting linguistic features of textual data in the plurality of online communications from the training data;

extracting contextual information from the training data, wherein the contextual information from the training data comprises purposes associated with the plurality of online communications, and the purposes comprise at least a first purpose and a second purpose; and

employing the training data, the linguistic features extracted from the training data, and the purposes indicated by the contextual information extracted from the training data to generate or update a formality classification model used to determine the formality level of the plurality of text items.

18. A system for recommending a text item to an online user, comprising:

memory storing computer program instructions; and

one or more processors that, in response to executing the computer program instructions, effectuate operations comprising:

machine-training a plurality of models based on parallel textual data including different versions of training sentences each having a same literal meaning but different formality levels;

receiving a request for recommending one or more text items of a plurality of text items to an online user;

extracting one or more linguistic features from each of the plurality of text items;

extracting contextual information with respect to the online user;

determining, using a first model of the plurality of models, a formality level of each of the plurality of text items based on the one or more linguistic features and the contextual information, wherein the formality level represents a degree of formality of the text item;

determining an identity of the online user based on the contextual information;

determining the formality preference of the online user based on a user profile associated with the identity;

ranking the plurality of text items based on their respective formality levels and the formality preference of the online user indicative of a degree of formality preferred by the online user;

selecting the one or more text items from the plurality of text items based on the ranking;

recommending the one or more text items to the online user; and

updating at least one of the plurality of models based on the one or more linguistic features and the contextual information.