Mining product aspects from opinion text

1. A computer implemented method for extracting aspects from a text stream, comprising:

receiving a text stream having one or more sentences;

parsing one or more of the sentences to determine corresponding subject-verb-object (SVO) triples comprising a subject of the sentence, an object of the sentence, and a verb in the sentence relating the subject to the object;

selecting at least one parsed sentence from the one or more parsed sentences having a corresponding SVO triple, the verb or a lemma of the verb of the SVO triple matching at least one of a predefined verb stored in a list of predefined verbs of interest;

identifying the subject of the selected at least one parsed sentence as an aspect candidate;

tokenizing and normalizing the identified aspect candidate;

generating one or more n-grams for the tokenized and normalized aspect candidate, wherein the n-gram comprises n characters, wherein n is a number varying from 1 up to and including 3;

determining, for the generated one or more n-grams, a frequency at which the one or more n-grams is generated;

selecting a subset of the generated one or more n-grams as aspects based on the frequency with which the one or more n-grams are generated; and

displaying the selected subset of the generated one or more n-grams as a visual display, wherein the generated one or more n-grams is displayed in a color to indicate the sentiment associated with the one or more n-grams, and wherein the visual display includes key phrases and review snippets associated with the aspect.

2. The method of claim 1 , wherein the normalizing includes lemmatization.

3. The method of claim 1 , further comprising filtering one or more of the identified aspect candidates and the selected aspects, based on a filtering criteria.

4. The method of claim 1 , wherein the text stream includes online product reviews.

5. The method of claim 1 , wherein the text stream includes posts on a discussion forum.

6. The method of claim 1 , wherein the text stream includes a set of digital text documents.

7. The method of claim 1 , wherein the text stream includes a set of social media posts.

8. A computer system for extracting aspects from a text stream, comprising:

a computer device having a processor and a tangible storage device, wherein the computer is attached to a first object; and

a program embodied on the storage device for execution by the processor, the program having a plurality of programming instruction including instructions to:

receive a text stream having one or more sentences;

parse one or more of the sentences to determine corresponding subject-verb-object (SVO) triples comprising a subject of the sentence, an object of the sentence, and a verb in the sentence relating the subject to the object;

select at least one parsed sentence from the one or more parsed sentences having a corresponding SVO triple, the verb or a lemma of the verb of the SVO triple matching at least one of a predefined verb stored in a list of predefined verbs of interest;

identify the subject of the selected at least one parsed sentence as an aspect candidate;

tokenize and normalized the identified aspect candidate;

generate one or more n-grams for the tokenized and normalized aspect candidate, wherein the n-gram comprises n characters, wherein n is a number varying from 1 up to and including 3;

determine, for the generated one or more n-grams, a frequency at which the one or more n-grams is generated;

select a subset of the generated one or more n-grams as aspects based on the frequency with which the one or more n-grams-are generated; and

display the selected subset of the generated one or more n-grams as a visual display, wherein the generated one or more n-grams is displayed in a color to indicate the sentiment associated with the one or more n-grams, and wherein the visual display includes key phrases and review snippets associated with the aspect.

9. The system claim 8 , wherein the normalizing includes lemmatization.

10. The system of claim 8 , further comprising filtering one or more of the identified aspect candidates and the selected aspects, based on a filtering criteria.

11. The system of claim 8 , wherein the text stream includes online product reviews.

12. The system of claim 8 , wherein the text stream includes posts on a discussion forum.

13. The system of claim 8 , wherein the text stream includes a set of digital text documents.

14. The system of claim 8 , wherein the text stream includes a set of social media posts.

15. A computer program product for extracting aspects from a text stream, comprising a non-transitory tangible storage device having program code embodied therewith, the program code executable by a processor of a computer to perform a method comprising:

receiving, by the processor, a text stream having one or more sentences;

parsing, by the processor, one or more of the sentences to determine corresponding subject-verb-object (SVO) triples comprising a subject of the sentence, an object of the sentence, and a verb in the sentence relating the subject to the object;

selecting, by the processor, at least one parsed sentence from the one or more parsed sentences having a corresponding SVO triple, the verb or a lemma of the verb of the SVO triple matching at least one of a predefined verb previously stored in a list of predefined verbs of interest;

identifying, by the processor, the subject of the selected at least one parsed sentence as an aspect candidate;

tokenizing and normalizing, by the processor, the identified aspect candidate;

generating, by the processor, one or more n-grams for the tokenized and normalized aspect candidate, wherein the n-gram comprises n characters, wherein n is a number varying from 1 up to and including 3;

determining, by the processor, for the generated one or more n-grams, a frequency at which the one or more n-grams is generated;

selecting, by the processor, a subset of the generated one or more n-grams as aspects based on the frequency with which the one or more n-grams are generated; and

displaying, by the processor, the selected subset of the generated one or more n-grams as a visual display, wherein the generated one or more n-grams is displayed in a color to indicate the sentiment associated with the one or more n-grams, and wherein the visual display includes key phrases and review snippets associated with the aspect.

16. The computer program product of claim 15 , wherein the normalizing includes lemmatization.

17. The computer program product of claim 15 , further comprising filtering, by the processor, one or more of the identified aspect candidates and the selected aspects, based on a filtering criteria.

18. The computer program product of claim 15 , wherein the text stream includes online product reviews.

19. The computer program product of claim 15 , wherein the text stream includes posts on a discussion forum.

20. The computer program product of claim 15 , wherein the text stream includes a set of digital text documents.