Topical sentiments in electronically stored communications
First Claim
Patent Images
1. A computer implemented method for obtaining topical sentiments from a plurality of electronically stored communications comprising:
- determining with the assistance of a computer whether each communication in a plurality of communications is topical to a first topic;
for each communication determined to be topical to the first topic, separating with the assistance of a computer the communication into least one expression;
for each of the at least one expression, determining with the assistance of a computer if the expression is topical to a second topic;
for each expression that is determined to be topical to the second topic, determining with the assistance of a computer a polarity of the expression; and
reporting the polarity of each expression.
5 Assignments
0 Petitions
Accused Products
Abstract
The present application presents methods for performing topical sentiment analysis on electronically stored communications employing fusion of polarity and topicality. The present application also provides methods for utilizing shallow NLP techniques to determine the polarity of an expression. The present application also provides a method for tuning a domain-specific polarity lexicon for use in the polarity determination. The present application also provides methods for computing a numeric metric of the aggregate opinion about some topic expressed in a set of expressions.
-
Citations
49 Claims
-
1. A computer implemented method for obtaining topical sentiments from a plurality of electronically stored communications comprising:
-
determining with the assistance of a computer whether each communication in a plurality of communications is topical to a first topic; for each communication determined to be topical to the first topic, separating with the assistance of a computer the communication into least one expression; for each of the at least one expression, determining with the assistance of a computer if the expression is topical to a second topic; for each expression that is determined to be topical to the second topic, determining with the assistance of a computer a polarity of the expression; and reporting the polarity of each expression. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
-
-
18. A method for calculating, from a plurality of electronically stored expressions, an aggregate metric which estimates a frequency of at least one of positive or negative polar expressions contained in the expressions, comprising:
-
determining with the assistance of a computer whether each communication in a plurality of communications is topical to a first topic; for each communication determined to be topical to the first topic, separating the communication into a plurality of electronically stored expressions; determining with the assistance of a computer an opinion contained in each of the plurality of electronically stored expressions based upon at least one of a presence, absence and strength of polarity associated with the first topic; calculating an aggregate metric from the determined opinions; and reporting the aggregate metric. - View Dependent Claims (19, 20, 21, 22, 23, 24, 25, 26, 27, 28)
-
-
29. A computer implemented method for finding at least one polar expression in a plurality of electronically stored communications, comprising:
-
determining with the assistance of a computer whether each communication in a plurality of communications is topical to a first predefined topic; for each communication determined to be topical to the first topic, separating the communication into at least one expression; for each expression, determining with the assistance of a computer if the expression is topical to a second topic; providing a polarity lexicon of at least one of sentimental words or sentimental phrases associated with a topical domain; for each expression that is determined to be topical to the second topic, utilizing with the assistance of a computer the polarity lexicon against the at least one word or phrase found in the expression; and assigning with the assistance of a computer a polarity associated with at least one matching sentimental word or sentimental phrase in the polarity lexicon to the at least one word or phrase in the expression. - View Dependent Claims (30, 31, 32, 33, 34)
-
-
35. A computer implemented method for obtaining topical sentiments from a body of electronically stored communications comprising:
-
isolating with the assistance of a computer a first subset of the communications topical to a particular topic; isolating with the assistance of a computer a second subset of communications from the first subset of communications where the second subset of communications includes polar segments; for each communication in the second subset of communications, separating the communication into at least one expression; for each of the at least one expression, determining with the assistance of a computer if the expression is topical to a second topic; and reporting each expression that is topical to the second topic. - View Dependent Claims (36, 37, 38, 39, 40, 41)
-
-
42. A computer implemented method for tuning a polarity lexicon for use in classifying polar expressions, comprising:
-
providing a polarity lexicon; determining with the assistance of a computer whether each communication in a plurality of communications is topical to a first topic; for each communication determined to be topical to the first topic, separating the communication into at least one expression; scanning the at least one expression for candidate words that have the potential to be added to the polarity lexicon; with the assistance of a computer implemented graphical user interface providing a user with the candidate words for at least one of addition, subtraction or exclusion to the polarity lexicon; and at least one of adding, subtracting or excluding each candidate word from the polarity lexicon according to input received by the graphical user interface. - View Dependent Claims (43, 44, 45, 46)
-
-
47. A computer implemented method for obtaining topical sentiments from a plurality of electronically stored communications comprising:
-
determining with the assistance of a computer whether each communication in a plurality of communications is topical to a first topic; for each communication determined to be topical to the first topic, determining with the assistance of a computer at least one topical expression in the communication; locating with the assistance of a computer one or more polar expressions in the topical expressions; and reporting an expression that is both a topical expression and a polar expression as containing a topical sentiment. - View Dependent Claims (48, 49)
-
Specification