System and method for detecting personal experience event reports from user generated internet content

Abstract

A method implementable on a computing device for detecting personal experience event reports from user generated content on the Internet is disclosed. The method includes filtering a collection of Internet posts to include only the Internet posts containing personal experience terms. The method additionally includes further filtering the filtered Internet posts by removing the Internet posts with non-personal experience terms.

What is claimed is: 
     
         1 . A method for detecting personal experience event reports from user generated content on the Internet, implementable on a computing device, the method comprising:
 filtering a collection of Internet posts to include only said Internet posts containing personal experience terms; and   further filtering said filtered Internet posts by removing said Internet posts with non-personal experience terms.   
     
     
         2 . A method according to  claim 1  and also comprising:
 compiling a list of post collection websites; and 
 collecting said Internet posts according to said list of websites for analyzing on a periodic basis. 
 
     
     
         3 . A method according to  claim 2  and wherein said compiling comprises at least one of:
 detecting “good” textual patterns indicative of an authentic user generated personal experience event report from a training set of authenticated user generated personal experience event reports; or 
 detecting “bad” textual patterns indicative of a non-authentic user generated personal experience event report from a training set of non-valid user generated personal experience event reports. 
 
     
     
         4 . A method according to  claim 3  and also comprising:
 assigning weights to each of said “good” and “bad” textual patterns to reflect a likelihood of said user generated personal experience event reports including each of said “good” and “bad” textual patterns 
 
     
     
         5 . A method according to  claim 4  and also comprising:
 assigning weights to predictive factors associated with said authentic and non-authentic user generated personal experience event reports in said training sets to reflect a likelihood of said user generated personal experience event reports being associated with at least some of said predictive factors, wherein said predictive factors include at least one of external website/page rankings and factors derived from said training sets. 
 
     
     
         6 . A method according to  claim 5  and wherein said derived factors include at least one of website metadata, number of images per page, number of links per page, ratio of authentic user generated personal experience event reports per discussion thread, number of authentic user generated product personal experience event reports per website, total anchor terms detected, and total terms detected. 
     
     
         7 . A method according to  claim 6  and also comprising
 identifying said candidate websites with Internet posts including terms from at least one of two “anchor” categories, wherein said anchor categories represent two essential components of user generated product personal experience reports; 
 collecting at least a sample of said Internet posts from said identified candidate websites; 
 scoring each candidate website according to a cumulative weighted score as per said set of weighted indicators, wherein a pre-defined score threshold indicates a website with user generated personal experience event reports; and 
 adding said website with said user generated personal experience event reports to said list of post collection websites. 
 
     
     
         8 . A method for compiling a list of Internet post collection websites, implementable on a computing device, the method comprising:
 detecting “good” textual patterns indicative of an authentic user generated personal experience event report from a training set of authenticated user generated personal experience event reports; and   detecting “bad” textual patterns indicative of a non-authentic user generated personal experience event report from a training set of non-valid user generated personal experience event reports.