Feature Set Extraction F5

PURPOSE

The following document describes the feature extraction process of feature set F5. Since the extraction code was written in Python the following example presents the Python code in text. In order to insert the Python coe it was put was a comment. Regular expressions were used to extract feature set F5. In order to protect the copyright of the provided REGEX the following code is only exemplary:

In case a given sentence matches the REGEX pattern, “TRUE” is returned for this sentence (see below)

Directives

\(Sentence\) MATCHES “(SAMPLE_REGULAR_EXPRESSION).*“=>TRUE

nrRegex = 7

Process

1. As a first step, a couple of functions were defined to check for the 7 regular expressions provided.

#def negative_match(line, regex):
    #print line
    #pattern = re.compile(regex)
    #m = pattern.match(line)
    #if m:
        #print ("negative match") #, m.groups()
        #return 1
    #else:
        #return 0

#def positive_match(line, regex):
    # type: (object, object) -> object
    #print line
    #pattern = re.compile(regex)
    #m = pattern.match(line)
    #if m:
        #print ("positive match") #, m.groups()
        #return 1
    #else:
        #return 0

2. Then, a function was written to check if any sentences matched the REGEX pattern:

#def match(line, totalResult):
    #result = [0] * nrRegex
    #result[0] = positive_match(line, r".*(SAMPLE REGEX).*")
    #result[1] = positive_match(line, r".*(SAMPLE REGEX).*")
    #result[2] = negative_match(line, r".*(SAMPLE REGEX).*")
    #result[3] = negative_match(line, r".*(SAMPLE REGEX).*")
    #result[4] = negative_match(line, r".*(SAMPLE REGEX).*")
    #result[5] = negative_match(line, r".*(SAMPLE REGEX).*")
    #result[6] = negative_match(line, r".*(SAMPLE REGEX).*")
    #for i in range(0,nrRegex,1):
        #totalResult[i] += result[i]
    #return (result, totalResult)

3. The end result was stored in a csv called “results.csv” containing nr’,‘id’,‘text’,‘positive directives 1’,‘positive directives 2’,‘negative directives 1’,‘negative directives 2’,‘negative directives 3’,‘negative directives 4’,’negative directives 5

#all_files = fm.load('resources') # fm.load('resources','w') will open in write mode
#all_files = ['file.csv']
#outputFile = 'result.csv'

#def readCSV(fileName, outputFile, totalResult):
    #with open(fileName, 'r') as csvfile:
        #readCSV = csv.reader(csvfile, delimiter=',')
        #for row in readCSV:
            #id   = row[0]
            #nr   = row[1]
            #text = row[2]
            #(result, totalResult) = match(text, totalResult)
            #outputFile.writerow((id, nr, "", result[0], result[1], result[2], result[3] , result[4], result[5],     result[6]))
    #return totalResult

#with open(outputFile, 'w') as csvOutFile:
    #writer = csv.writer(csvOutFile, delimiter=',', quotechar='"', quoting=csv.QUOTE_NONNUMERIC)
    #writer.writerow(('nr','id','text','p1','p2','n1','n2','n3','n4','n5'))
    #totalResult = [0] * nrRegex
    #for f in all_files:
        #if (-1 == f.find(".csv")):
            #readNonCSV(f, writer, totalResult)
        #else:
            #totalResult = readCSV('resources/' + f, writer, totalResult)
    #write the total results
   # writer.writerow(('-1', '-1', '', totalResult[0], totalResult[1], totalResult[2], totalResult[3], totalResult[4], #totalResult[5], totalResult[6]))