A Comparison of the Vector Space Model Method and Winnowing Algorithm to Measure the Similarity of Documents

Eva Y Puspaningrum (Informatics, Universitas Pembangunan Nasional “Veteran” Jawa Timur, Jl.Rungkut Madya no.01, Surabaya, Indonesia)
Budi Nugroho (Informatics, Universitas Pembangunan Nasional “Veteran” Jawa Timur, Jl.Rungkut Madya no.01, Surabaya, Indonesia)
Firza Prima A (Informatics, Universitas Pembangunan Nasional “Veteran” Jawa Timur, Jl.Rungkut Madya no.01, Surabaya, Indonesia)

Proceeding International Conference on Information Technology and Business
2019: INTERNATIONAL CONFERENCE ON INFORMATION TECHNOLOGY AND BUSINESS (ICITB) 5

The growth of information and communication technology has increased significantly from year to year. The issue that is developing now is the number of documents that are copied and paste. The amount of text data is constantly increasing in cyberspace so that everyone can easily find the documents they need. Because of these problems, measuring the similarity of the two documents is necessary and is fundamental to detecting plagiarism from many different documents. In this work, we would like to compare the effectiveness of the algorithm used to measure the similarity between two documents. Winnowing and SVM algorithms are widely used to compare documents because the plot is easy to understand and easy to use. The Experiment Result, we can find that the performance of fingerprints and winnowing is better than VSM. Moreover, the winnowing algorithm is more stable than others.Keywords: Vector Space Model, Winnowing, Similarity of Documents