Reconciling detailed transaction feedback

• US 9,779,428 B2
• Filed: 10/19/2016
• Issued: 10/03/2017
• Est. Priority Date: 04/12/2013
• Status: Active Grant

• Current Assignee: eBay Inc.
• Original Assignee: eBay Inc.
• Inventors: Maddali, Lakshman Shyam Sundar, Sharma, Avani Goel
• Primary Examiner(s): YEN, ERIC L
• Application Number:
  US15/297,244
• Publication Number:
  US 20170039606A1
• Time in Patent Office: 349 Days
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/2465 Query processing support for facilitating data mining operations in structured databases
• CPC Class Codes: G06F 40/279 Recognition of textual entities
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06Q 30/0282 Rating or review of business operators or products

Abstract

Reconciling detailed transaction feedback by detecting a rating of a transaction, where the rating indicates a negative experience, mining the sentiment of words in feedback text that is included with or as part of the rating to detect whether the words indicate positive sentiment or negative sentiment, responsive to determining that the words in the feedback text indicate that the feedback text connotes a positive sentiment, adjusting the rating of the transaction. The mining may include testing words in the feedback text to detect whether the words indicate positive sentiment or negative sentiment by calculating a sentiment score.

Claims

1. A method for determining the sentiment of a plurality of words as negative or positive, comprising:

receiving, by one or more computer processors, a plurality of words from a user interface of a client machine;

for each word in the plurality of words, testing whether the word indicates positive or negative sentiment by using a sentiment mining tool;

responsive to detecting an indication of positive sentiment of the word, testing whether there is, in the plurality of words, a known negative word that was mined from the text of feedback from one or more ecommerce transaction logs;

responsive to detecting that there is, in the plurality of words, no known negative word that was mined from the text of feedback from one or more ecommerce transaction logs, incrementing a positive word counter in one or more hardware processors;

responsive to detecting an indication of negative sentiment of the word, testing whether the word is in a list of negative modifiers; and

responsive to detecting that the word is not in the list of negative modifiers, incrementing a negative word counter in the one or more hardware processors.

2. The method of claim 1, the method further comprising:

responsive to detecting that all words in the plurality of words have been tested, determining a numerical quantity from a count of the positive word counter and a count of the negative word counter; and

testing the numerical quantity and, responsive to detecting the numerical quantity as being equal to or greater than a threshold, marking the plurality of words as positive sentiment, and, responsive to detecting the numerical quantity as being less than the threshold, marking the plurality of words as negative.

3. The method of claim 1 wherein testing whether the word indicates positive sentiment comprises detecting whether the word is a word that is in a preselected positive lexicon.

4. The method of claim 1 wherein testing whether the word that indicates negative sentiment comprises detecting whether the word is one of a negative modifier word, a contrasting conjunction word, or a word that is known to have a negative connotation.

5. The method of claim 1 wherein responsive to detecting that there is a known negative word in the plurality of words, flagging the plurality of words as indicating a negative sentiment.

6. The method of claim 1 wherein responsive to detecting that the word is in the list of negative modifiers, flagging the plurality of words as indicating a negative sentiment.

7. The method of claim 1 wherein the plurality of words comprises a second group of words formed by testing whether there is a punctuation mark in a first group of words and, responsive to detecting that there is no punctuation mark in the first group of words, using a probabilistic technique to break the first group of words into pluralities of words.

8. One or more computer-readable hardware storage device having stored therein a set of instructions which, when executed by one or more processors of a computer, causes the computer to execute operations comprising:

receiving a plurality of words from a user interface of a client machine;

for each word in the plurality of words, testing whether the word indicates positive or negative sentiment by using a sentiment mining tool;

responsive to detecting an indication of positive sentiment of the word, testing whether there is, in the plurality of words, a known negative word that was mined from the text of feedback from one or more ecommerce transaction logs;

responsive to detecting that there is, in the plurality of words, no known negative word in the plurality of words that was mined from the text of feedback from one or more ecommerce transaction logs, incrementing a positive word counter;

responsive to detecting an indication of negative sentiment of the word, testing whether the word is in a list of negative modifiers; and

responsive to detecting that the word is not in the list of negative modifiers, incrementing a negative word counter.

9. The one or more computer-readable storage device of claim 8, the operations further comprising:

responsive to detecting that all words in the plurality of words have been tested, determining a numerical quantity from a count of the positive word counter and a count of the negative word counter; and

testing the numerical quantity and, responsive to detecting the numerical quantity as being equal to or greater than a threshold, marking the plurality of words as positive sentiment, and, responsive to detecting the numerical quantity as being less than the threshold, marking the plurality of words as negative.

10. The one or more computer-readable storage device of claim 8 wherein testing whether the word indicates positive sentiment comprises detecting whether the word is a word that is in a preselected positive lexicon.

11. The one or more computer-readable storage device of claim 8 wherein testing whether the word indicates negative sentiment comprises testing whether the word is one of a negative modifier word, a contrasting conjunction word, or a word that is known to have a negative connotation.

12. The one or more computer-readable storage device of claim 8 wherein responsive to detecting that there is a known negative word in the plurality of words, flagging the plurality of words as indicating a negative sentiment.

13. The one or more computer-readable storage device of claim 8 wherein responsive to detecting the word is in the list of negative modifiers, flagging the plurality of words as indicating a negative sentiment.

14. The one or more computer-readable storage device of claim 8 wherein the plurality of words comprises a second group of words formed by testing whether there is a punctuation mark in a first group of words and, responsive to detection that there is no punctuation mark in the first group of words, using a probabilistic technique to break the first group of words into pluralities of words.

15. A system for determining the sentiment of a plurality of words as negative or positive, comprising:

one or more hardware processors and storage configured to

receive a plurality of words from a user interface of a client machine;

for each word in the plurality of words, test whether the word indicates positive or negative sentiment by use of a sentiment mining tool;

responsive to detection an indication of positive sentiment of the word, test whether there is, in the plurality of words, a known negative word that was mined from the text of feedback from one or more ecommerce transaction logs;

responsive to detection that there is, in the plurality of words, no known negative word that was mined from the text of feedback from one or more ecommerce transaction logs, increment a positive word counter;

responsive to detection of an indication of negative sentiment of the word, test whether the word is in a list of negative modifiers; and

responsive to detection that the word is not in the list of negative modifiers, increment a negative word counter.

16. The system of claim 15 the one or more hardware processors and storage further configured:

responsive to detection that all words in the plurality of words have been tested, determine a numerical quantity from a count of the positive word counter and a count of the negative word counter; and

test the numerical quantity and, responsive to detection of the numerical quantity as being equal to or greater than a threshold, mark the plurality of words as positive sentiment, and, responsive to detection of the numerical quantity as being less than the threshold, mark the plurality of words as negative.

17. The system of claim 15 wherein the test for whether the word indicates positive sentiment comprises a detection of whether the word is a word that is in a preselected positive lexicon.

18. The system of claim 15 wherein the test for whether the word indicates negative sentiment comprises detection of whether the word is one of a negative modifier word, a contrasting conjunction word, or a word that is known to have a negative connotation.

19. The system of claim 15 wherein responsive to detection that there is a known negative word in the plurality of words, flag the plurality of words as indicating a negative sentiment.