Text Summarization using TextRank for Knowledge Externalization from Indonesian Online Discussion Forums

Online discussion forum is one of digital media that contain many opinions about some issue. This research aims to use automatic text summarization technology for knowledge externalization from online discussion forums in Indonesia. The method that used in this research is the TextRank algorithm which is in summarizing the text gives quite good results. It is proven based on experiment and evaluation result using ROUGE-L evaluation to evaluate the similarity between manual summary and output summary from system. The experiment used 5 Indonesian Islamic online discussion forums. The summary results of 20 documents as sample data and 31 students carry out testing show that summary results are easy to understand. The test results obtained that the overall average value of Recall is 0.42 and Precision is 0.42. Simultaneously, the overall average F-score is 0.39 with a maximum F-score of 0.72 and a minimum F-score of 0.052. While the test for user understanding of the summary text results obtained, 44.52% very understand, 34.19% understand, and 16.13% understand enough. Meanwhile, students who stated that they did not understand the results of automatic text summary 2.26% and 0.97% stated that they did not understand.