Citation Detection on Scientific Journal Using Support Vector Machine

Scientific journals are scientific works that are published regularly by an organization or institution, writing carelessness in scientific work can be considered as a form of plagiarism. So writing quotations in scientific work is important, because quotations are able to provide recognition of reference sources. In finding out the source of the citation, the method used in this study is SVM. In the process, two experiments were made, including the first experiment using a combination of SVM and TF-IDF, where the TF-IDF method was used to reduce the number of dimensions of the data, so that data could be processed optimally using SVM, while the second experiment used SVM only. The results of the study for the first experiment obtained accuracy and f-measure values of 98.79% and 98.44%. While the second experiment obtained accuracy and f-measure values of 98.1% and 97.64%. From these results it can be seen that the use of the TF-IDF method is able to support SVM as evidenced by the value of accuracy and f-measure which has increased by 0.69% and 0.8%.