Enhancing Target Document Search in CopyCatch: A Focus on Thai and English document

CopyCatch, is an automatic software designed to detect plagiarism in electronic documents written in both Thai and English. It consists of two main components: Target Document Search and Document Comparison and Similarity Calculation. The Target Document Search is an important step in identifying a list of target documents from the document corpus that are expected to have content similar to the document being examined. In this paper, we propose a novel approach to building the document corpus and comparing input documents with target documents by focusing on Thai vowels and English alphabets. The comparison process involves controlling three variables: passage size, window size, overlap of 3-gram chunks, and the frequency matching of 3-gram chunk in the document corpus. We found that setting the passage size to 10,000 characters, using a window size and overlap of 10:5 for 3-gram chunk, and applying a frequency matching condition of 1 had a substantial impact on the accuracy and number of files identified as similar.