Implementasi Metode Ekstraksi Textrank Dan Agglomerative Hierarchical Clustering Untuk Pengelompokkan Jurnal Berdasarkan Abstrak Berbasis Website

Abstract

A digital library is a library that uses information technology and its collections in digital form, can be accessed anytime and anywhere and the dissemination of information is very fast, precise, and accurate. In digital libraries, there are various types of scientific writings (journals). To facilitate the search process from the user, the journals need to be grouped according to certain criteria. To solve these problems, a classification method can be applied. In the process, classification can be done manually or with the help of technology. Manually, classification is done by humans without any help from computer intelligent algorithms. However, this manual process is time consuming and inefficient. Classification can be done with the help of technology, one of which is by using the Agglomerative Hierarchical Clustering algorithm. To simplify the classification process, it is necessary to summarize the set of texts in the document into a few important sentences. In this study, the TextRank method will be used. One of the advantages of this algorithm is that there is no need for training using training data on the algorithm used. The way TextRank works is to find the sentence that is most similar to all the sentences in the text. The sentence that is most similar to all the sentences will be the most important sentence in the text. The result of this research is a website that can be used to group journals. From the results of the tests carried out, the website is able to carry out the journal classification process with a success rate of 93.33%.