Paragraph Selection Methods Using Feature-Based on Segment-Based Clustering Process Using Paragraphs for Identifying Topics on Indications Detection of Plagiarism System

In segment-based clustering, the paragraphs selection as a data-set in the clustering process has a very important role. This is because the paragraph used as the data-set can affect the clustering result. In this research used the method of paragraph selection using feature-based, which aims to optimize the clustering process done in previous research. Based on evaluating using Silhouette Coefficient and Sum Square Errors evaluation methods to find the proper k value, it was found that with the use of Feature Based method, there was better result compared to the evaluation result from previous research.