Classification of user sentiment based on machine learning

1. A system for machine learning classification of user sentiment, the system comprising:

a sentiment analysis module configured for communication with a search interface, the sentiment analysis module comprising:

a data store including a plurality of category information, wherein the plurality of category information includes a set of domain-specific category information;

one or more extraction engines configured to extract a plurality of aspects from textual data obtained from the search interface,

wherein the plurality of aspects comprises implicit aspects and explicit aspects; and

a machine learning model configured to:

receive the plurality of aspects and the set of domain-specific category information; and

generate a sentiment based on the plurality of aspects and the set of domain-specific category information;

wherein the search interface is configured to:

provide user inputs entered through the search interface to the sentiment analysis module;

receive the sentiment from the sentiment analysis module; and

update the plurality of category information based on the sentiment and the plurality of aspects; and

wherein the one or more extraction engines are configured to extract the plurality of aspects from textual data by:

clustering one or more words or one or more phrases from the textual data,

determining a most frequent word or phrase in a cluster,

labeling the most frequent word or phrase as an explicit aspect,

determining a part of speech associated with the one or more words or one or more phrases,

responsive to the part of speech, determining a set of synonyms and antonyms of the one or more words or one or more phrases associated with an adjective part of speech,

responsive to the part of speech, lemmatizing noun forms of the set of synonyms and antonyms to generate a lemmatized noun form,

determining a match of the lemmatized noun form with the explicit aspects, and

selecting the explicit aspects based on the match having a greatest number of occurrences to represent an aspect of an implicit cluster.

2. The system of claim 1 , wherein the machine learning model comprises:

a gate mechanism configured to regulate the plurality of category information flow in the machine learning model, wherein the gate mechanism filters a portion of the plurality of category information and passes the set of domain-specific category information.

3. The system of claim 2 , wherein the gate mechanism is further configured to filter the set of domain-specific category information and pass a subset of domain-specific category information.

4. A method for machine learning classification of user sentiment, the method comprising:

storing a plurality of category information, wherein the plurality of category information includes a set of domain-specific category information;

extracting a plurality of aspects from textual data obtained from a search interface, the extracting comprising:

clustering one or more words or one or more phrases from the textual data,

determining from the textual data, a most frequent word or phrase in a cluster as a meaning of the cluster,

labeling the most frequent word or phrase as an explicit aspect,

determining a part of speech associated with the one or more words or one or more phrases,

responsive to the part of speech, determining a set of synonyms and antonyms of the one or more words or one or more phrases associated with an adjective part of speech,

responsive to the part of speech, lemmatizing noun forms of the set of synonyms and antonyms to generate a lemmatized noun form,

determining a match of the lemmatized noun form with explicit aspects, and

selecting the explicit aspect based on the match having a greatest number of occurrences to represent the aspect of an implicit cluster; and

generating a sentiment by a machine learning model, the generating comprising:

receiving the plurality of aspects and the set of domain-specific category information; and

generating a sentiment based on the plurality of aspects and the set of domain-specific category information;

wherein the plurality of aspects comprises an implicit aspect and the explicit aspect.

5. The method of claim 4 , wherein the machine learning model comprises:

a gate mechanism configured to regulate the plurality of category information flow in the machine learning model, wherein the gate mechanism filters a portion of the plurality of category information and passes the set of domain-specific category information.

6. The method of claim 4 , wherein gate mechanism is further configured to filter the set of domain-specific category information and pass a subset of domain-specific category information.

7. The method of claim 4 further comprising:

receiving, by the search interface, the sentiment from a sentiment analysis module; and

updating, by the search interface, the plurality of category information based on the sentiment and the plurality of aspects.

8. A method of for gated machine learning, the method comprising

storing a plurality of category information, wherein the plurality of category information includes a set of domain-specific category information;

extracting a plurality of aspects from textual data obtained from a search interface, the extracting comprising:

determining a part of speech associated with one or more words or one or more phrases,

responsive to the part of speech, determining a set of synonyms and antonyms of the one or more words or one or more phrases associated with an adjective part of speech,

responsive to the adjective part of speech, lemmatizing noun forms of the set of synonyms and antonyms to generate a lemmatized noun form,

determining a match of the lemmatized noun form with explicit aspects, and

selecting the explicit aspects based on the match having a greatest number of occurrences to represent an aspect of an implicit cluster; and

generating a sentiment by a machine learning model, the generating comprising:

receiving the plurality of aspects and the set of domain-specific category information; and

generating a sentiment based on the plurality of aspects and the set of domain-specific category information;

filtering, by a gate mechanism, the plurality of category information in the machine learning model, wherein the gate mechanism filters the set of domain-specific category information, wherein the gate mechanism comprises applying a weight to the domain-specific category information;

updating the plurality of category information based on the sentiment and the plurality of aspects; and

updating the weight of the gate mechanism based on the sentiment and the plurality of aspects.

9. The method of claim 8 , wherein gate mechanism is further configured pass a subset of the set of domain-specific category information.

10. The method of claim 8 , wherein the plurality of aspects comprises an implicit aspect and an explicit aspect.

11. The method of claim 8 , wherein the extracting comprises:

clustering one or more words or one or more phrases from the textual data;

determining from the textual data, a most frequent word or phrase in a cluster as a meaning of the cluster; and

labeling the most frequent word or phrase as an explicit aspect.

12. The method of claim 8 , wherein updating the plurality of category information based on the sentiment and the plurality of aspects comprises adding a new aspect to the plurality of aspects based on the sentiment exceeding a threshold.

13. The method of claim 8 , wherein updating the weight of the gate mechanism based on the sentiment and the plurality of aspects comprises adjusting a value of the weight responsive to the sentiment exceeding a threshold.