Skip to main content
. 2023 Aug 29;25(9):1271. doi: 10.3390/e25091271
Algorithm 2: QGA for Plagiarism Detection
Input: Dataset Xsrc; Suspicious Document Xsusp; QGA Parameters, WordNet
  • 1-

    while n < size of documents do

  • 2-

      S ← Sentence Segmentation (Xsrc)

  • 3-

      y ←0

  • 4-

      While y < S ! = NULL do

  • 5-

      T ← Tokenization (S)

  • 6-

      z←0

  • 7-

      while z < size of T do

  • 8-

       M ← POS Tagging (T)

  • 9-

       N ← Lemmatization (M)

  • 10-

       z++

  • 11-

      end

  • 12-

      tf-isf (N)

  • 13-

      y++

  • 14-

     end

  • 15-

     n++

  • 16-

    end

  • 17-

    t ← 0

  • 18-

    while termination condition not satisfied do

  • 19-

      t ← t+1

  • 20-

      Call Algorithm 1   // QGA Procedure

  • 21-

      Return Best_Pop ←New_Pop // Store the best solution among P(t)

  • 22-

    end

  • 23-
    sim1← sum of words in Xsusp // the number of common word-level concepts in Xsusp
    •     that collects nouns and verbs
  • 24-

    sim2← sum of words in Xsrc // the number of common word-level concepts in Xsrc

  • 25-

    If sim1−sim2>ε

  • 26-

     Doc. Status = =Plagiarized

  • 27-

    end

  • 28-

    For each suspicious-source word pair (wq,wk) //To compute the semantic similarity

  • 29-

       - WordNet is used to derive the synset lists Wq_syn ofwq and Wk_synof wk

  • 30-

       of each word.

  • 31-

       - Only synsets in the same POS class as the word are retrieved for these lists

  • 32-

    end

  • 33-
    Count ←The common words between the compared suspicious-source sentence
    • pair(Ssusp,Ssrc_sel) // Ssrc_sel is the best set of selected source sentences
    •    extracted from QGA’ procedure
  • 34-

    If count > τ

  • 35-

     Doc. Status = = Plagiarized

  • 36-

    end

  • 37-

    Else

  • 38-

     Doc. Status = = not plagiarized

  • 39-

    end

  • 40-

    Output = Doc. Status