Large-scale sentiment analysis

Abstract

A method for determining a sentiment associated with an entity includes inputting a plurality of texts associated with the entity, labeling seed words in the plurality of texts as positive or negative, determining a score estimate for the plurality of words based on the labeling, re-enumerating paths of the plurality of words and determining a number of sentiment alternations, determining a final score for the plurality of words using only paths whose number of alternations is within a threshold, converting the final scores to corresponding z-scores for each of the plurality of words, and outputting the sentiment associated with the entity.

1 - 7 . (canceled) 
     
     
         8 . A method performed by a specifically programmed computer system for determining a statistical sentiment associated with an entity, the method comprising:
 inputting a plurality of texts associated with the entity;   formatting the plurality of texts;   processing the plurality of texts using a sentiment lexicon;   determining, using the specifically programmed computer system, an entity statistical sentiment for the plurality of texts processed based on terms in the sentiment lexicon which are associated with text corresponding to the entity in the plurality of texts processed;   determining a world statistical sentiment based on terms in the sentiment lexicon in the plurality of texts processed;   normalizing the entity statistical sentiment based on the world statistical sentiment to obtain a normalized entity statistical sentiment; and   outputting the normalized entity statistical sentiment.   
     
     
         9 . The method of  claim 8 , further comprising determining a sentiment index based on a rank of the normalized entity statistic sentiment compared to normalized statistic sentiments of other entities. 
     
     
         10 . The method of  claim 8 , wherein processing the plurality of texts using the sentiment lexicon further comprises:
 identifying a position of text corresponding to the entity in the text;   identifying a position of terms in the sentiment lexicon in the text;   determining a polarity measure of the entity; and   determining a subjectivity measure of the entity.   
     
     
         11 . The method of  claim 10 , wherein determining the polarity measure of the entity further comprises associating positive and negative sentiment references using the terms in the sentiment lexicon with the entity. 
     
     
         12 . The method of  claim 10 , wherein determining the subjectivity measure of the entity further comprises accumulating counts of positive and negative sentiment references of the entity. 
     
     
         13 . The method of  claim 8 , further comprising translating at least one text of the plurality of texts into a target language. 
     
     
         14 . The method of  claim 8 , wherein the normalized entity statistical sentiment is output in terms of a comparison to another entity. 
     
     
         15 . The method of  claim 14 , wherein the comparison is a percentile rank. 
     
     
         16 . The method of  claim 8 , further comprising identifying and eliminating duplicate texts. 
     
     
         17 . The method of  claim 8 , further comprising:
 processing the plurality of texts using at least a first sentiment lexicon and a second sentiment lexicon, different from the first sentiment lexicon;   determining a first entity statistical sentiment for the plurality of texts processed based on terms in the first sentiment lexicon which are associated with text corresponding to the entity in the plurality of texts processed; and   determining a second entity statistical sentiment for the plurality of texts processed based on terms in the second sentiment lexicon which are associated with text corresponding to the entity in the plurality of texts processed.   
     
     
         18 . A computer system configured to determine a statistical sentiment associated with an entity, the computer system comprising a memory and a processor and being configured to:
 input a plurality of texts associated with the entity;   format the plurality of texts;   process the plurality of texts using a sentiment lexicon;   determine an entity statistical sentiment for the plurality of texts processed based on terms in the sentiment lexicon which are associated with text corresponding to the entity in the plurality of texts processed;   determine a world statistical sentiment based on terms in the sentiment lexicon in the plurality of texts processed;   normalize the entity statistical sentiment based on the world statistical sentiment to obtain a normalized entity statistical sentiment; and   output the normalized entity statistical sentiment.