An Improvement of Plagiarized Area Detection System Using Jaccard Correlation Coefficient Distance Algorithm

摘要

In this paper, a plagiarized area detection system is proposed in which Jaccard correlation coefficient is used for filtering to improve the processing time against huge volume of documents. Hence, the proposed system does filter to efficiently detect plagiarized area against huge volume of original documents by two algorithms; Jaccard coefficient distance algorithm and Cosine distance algorithm. Since Jaccard coefficient distance algorithm computes the distance between two document based only on the existence of words while Cosine distance algorithm uses word's frequency also, Jaccard coefficient distance algorithm is faster than Cosine one. Hence, for the efficiency, we use Jaccard coefficient distance algorithm as the first filter. According to the experiment result of the performance comparison between the proposed system and the previous our system, the newly proposed system outperforms the previous one with about 30% reduced processing time.