fixed some pattern detection bugs

issuecomp-analysis.R

@@ -9,8 +9,11 @@ date_start <- as.Date("2014-01-01")
 date_end <- as.Date("2014-12-31")
 drange <- as.integer(date_end - date_start)
 drange <- date_start + days(0:drange)
-issues <- data.frame(date = drange)
 
+
+# MATCH TWEETS ------------------------------------------------------------
+
+issues <- data.frame(date = drange)
 issuelist <- xmlToList("issues.xml")
 issueheads <- names(issuelist)
 issues[issueheads] <- 0
@@ -28,44 +31,38 @@ for(d in 1:nrow(issues)) {
     curtext <- as.character(tweets_curday$text[t])
     curtext <- str_replace_all(curtext, "#", "")
     
+    # Now test each single issue (not tag!)
     for(i in 1:length(issuelist)) {
       curtags <- as.character(issuelist[[i]])
       curissue <- names(issuelist)[i]
-      curtags <- str_c("\\W", curtags, "\\W")
-      tags_found <- str_detect(curtext, sprintf("%s", curtags))
-      tags_found <- any(tags_found)
-      
-      ######
-      
-      # Test all tags in ONE issue
+
+      # Now test all tags of a single issue
       for(t in 1:length(curtags)) {
-        curtag <- curtags[t]
-        curchars <- nchar(curtag, type = "chars")
+        curtag <- str_c("\\W", curtags[t], "\\W")
+        curchars <- nchar(curtag, type = "chars") - 4
         
+        # Match current tweet with tag. If >= 5 letters allow 1 changed letter, if >=8 letters allow 2 (Levenshtein distance)
         tags_found <- smartPatternMatch(curtext, curtag, curchars)
-        
         if(tags_found == 1) {
-          cat("Text contains at least the tag:", curtag, "\n")
+          #cat("Matched", curtag, "with", curtext,"\n")
+          issues[d,curissue] <- issues[d,curissue] + 1
           break
         }
-      }
-      
-      
-      ######
-      
-      if(tags_found) {
-        #cat("Positive in", curissue,"from",as.character(drange[d]),"\n")
-        issues[d,curissue] <- issues[d,curissue] + 1
-      }
-      else {
-        #cat("Nothing found\n")
-      }
+        else {
+          #cat("Nothing found\n")
+        }
+      } # /for curtags
       
     } # /for issuelist
   } # /for tweets_curday
 } # /for drange
 
 
+
+# WEEKLY INTERVALS --------------------------------------------------------
+
+
+
 ## Do not use days but week intervals
 
 wrange <- (as.integer(date_end - date_start) / 7)